DE102018004575A1 - Method for monitoring a chassis of a vehicle - Google Patents

Abstract

The invention relates to a method (V) for monitoring a chassis of a vehicle.
According to the invention, an actual chassis behavior is compared with an optimum chassis behavior learned by machine learning, and a deviation of the actual chassis behavior from the learned optimum suspension behavior is attributed to chassis wear and / or chassis damage, in which case corresponding information (I) is generated by means of at least one output unit (6). is output of the vehicle and / or a corresponding information (I) to a vehicle external device (7) is transmitted and / or a suspension setting of the designed as an adaptive chassis chassis is changed so that the actual suspension behavior is adapted to the learned optimum suspension behavior.

Description

The invention relates to a method for monitoring a chassis of a vehicle.

From the prior art, as in the DE 10 2009 001 689 A1 described a method for determining a state of a chassis component and properties of the road on the basis of a, a force on a steering arrangement assignable, issued by a sensor signal known. The determination is made by picking up at least one signal output from the sensor and attributable to the force, evaluating the picked-up signal by a processor and generating an output signal by an output device based on the evaluated signal, the output signal corresponding to the condition of the chassis component.

The invention is based on the object to provide a comparison with the prior art improved method for monitoring a chassis of a vehicle.

The object is achieved by a method for monitoring a chassis of a vehicle with the features of claim 1.

Advantageous embodiments of the invention are the subject of the dependent claims.

In a method for, in particular continuous, monitoring of a running gear of a vehicle, according to the invention an actual running gear behavior is compared with an optimal running gear learned by machine learning. Ie. For example, during the development phase, test drives are carried out with one or more vehicles in which the chassis is in an optimal state. This is the, in particular vehicle-specific, suspension behavior, which is thus an optimal suspension behavior, which was set and tested in a known manner by means of appropriate classical development methods, learned by machine learning in a control unit. Ie. a method of machine learning is trained on the optimal suspension behavior. The optimal suspension behavior learned in this way is then compared, in particular during a normal driving operation of the vehicle, with the actual suspension behavior that occurs.

If these are identical or substantially identical, the chassis of the vehicle has a good condition, in particular no chassis wear and no chassis damage. If, however, the actual suspension behavior deviates from this learned optimum suspension behavior, then due to this deviation, a chassis wear and / or chassis damage is inferred.

If this is the case, d. H. If a deviation of the actual chassis behavior from the learned optimal suspension behavior is determined and therefore concluded on a chassis wear and / or chassis damage, then a corresponding information by means of at least one output unit of the vehicle, for example, a so-called head unit, d. H. a central output unit, in particular display unit, of the vehicle, issued, and / or such a corresponding information is transmitted to a vehicle external device, in particular a manufacturer of the vehicle, in particular to a data center of the manufacturer of the vehicle.

If the chassis of the vehicle is designed as an adaptive chassis, it can alternatively or additionally be provided that a chassis setting of this chassis is changed, in particular optimized, in such a way that the actual chassis behavior is adapted to the learned optimum suspension behavior.

The term "suspension behavior" as used herein should, alternatively or additionally, also be understood to mean a chassis pattern and / or chassis behavior pattern.

In particular, a chassis wear, for example, worn shock absorbers, a game in a steering of the vehicle, an obscured lane and / or a defect in a ball joint, a control arm and / or a stabilizer, occurs creeping and therefore usually unnoticed by a driver of the vehicle. The driving behavior that changes over a longer period of time is unconsciously compensated by the driver through tax corrections.

By means of the solution according to the invention, such a chassis wear or, for example, an occurring chassis damage is recognized in good time and advantageously the driver is then informed so that he can initiate appropriate measures, in particular repair measures. As a result, for example, accidents can be avoided that may occur due to the chassis wear or chassis damage.

For example, worn and / or defective shock absorbers, more precisely vibration dampers lead to an extended braking distance, to increased aquaplaning Reduction of the effectiveness of an electronic stability program and a worse cornering. This can be avoided by means of the solution according to the invention.

If the vehicle has an adaptive chassis, by means of the solution according to the invention, a high level of ride comfort and good chassis properties can be maintained even in the event of occurring chassis wear and / or chassis damage by automatically optimizing the adaptive chassis and driving safety can be increased.

Since the chassis behavior is automatically learned by machine learning and can be quickly learned to other vehicle types and / or equipment variants, also a cost rolling out of the process and the technical system for its implementation is possible, d. H. a cost-effective implementation in the respective vehicle type and / or the respective equipment variant.

Embodiments of the invention are explained in more detail below with reference to drawings.

• 1 schematisch eine Vorrichtung zur Durchführung eines Verfahrens zur Überwachung eines Fahrwerks eines Fahrzeugs,
• 2 schematisch ein Ablaufdiagramm zum Lernen eines optimalen Fahrwerksverhaltens,
• 3 schematisch ein Ablaufdiagramm eines Verfahrens zur Überwachung eines Fahrwerks eines Fahrzeugs, und
• 4 schematisch ein Ablaufdiagramm zum Optimieren von Fahrwerkseinstellungen.

Showing:

• 1 1 schematically shows an apparatus for carrying out a method for monitoring a chassis of a vehicle,
• 2 schematically a flow chart for learning an optimal suspension behavior,
• 3 schematically a flowchart of a method for monitoring a chassis of a vehicle, and
• 4 schematically a flow chart for optimizing suspension settings.

Corresponding parts are provided in all figures with the same reference numerals.

Based on 1 to 4 Below is a procedure V for, in particular continuous, monitoring a chassis of a vehicle described, whereby a, in particular continuous, checking and ensuring a chassis functionality of the chassis of the vehicle is made possible.

For this purpose, during a development of the vehicle, the vehicle-specific chassis behavior by means of machine learning in a control unit 4 learned, ie it is trained by means of a method of machine learning, for example by means of regression methods to predict from steering movements the vehicle behavior and thus the chassis behavior, and / or by means of recurrent neural networks to map steering behavior. Due to the optimally adjusted and neither worn nor damaged chassis, the optimum chassis behavior is thus learned.

The term "suspension behavior" used here and below is intended to mean, alternatively or additionally, also a chassis pattern and / or chassis behavior pattern.

The chassis model learned in this way, d. H. the suspension behavior, for example, may be contextual, for example, based on a respective outside temperature and / or a respective load. It can thus be provided that for different outside temperatures and / or loads and / or other context criteria different, for each context optimal, suspension behavior are learned.

The learning of the optimal suspension behavior, for example online, d. H. during test drives, and / or offline, d. H. using collected data from test drives.

In the process V In order to monitor the chassis of the vehicle, a respective actual chassis behavior of the vehicle is then compared with the learned optimum suspension behavior for a respective customer, ie after delivery of the vehicle to the respective customer and thus in normal vehicle operation, in particular over a relatively long period of time. for example, to compensate for environmental influences and loading conditions, if this is not already covered by the context and the use of the appropriate context-related optimal chassis behavior. As soon as a deviation of the actual suspension behavior from the learned during the vehicle development, ie trained, optimal suspension behavior is determined, advantageously the customer, ie in particular the driver of the vehicle, informed by providing appropriate information I by means of at least one output unit 6 of the vehicle, in particular a head unit, is output.

Alternatively or additionally, for example, be provided, this corresponding information I to an off-vehicle device 7 to be transmitted, in particular to an off-board device 7 a manufacturer of the vehicle, for example to a data center of the manufacturer of the vehicle. This can, for example, information I about a chassis wear and chassis damage in the field, ie delivered at customer vehicles in a normal driving, are aggregated and, for example, can respective customers, in particular the driver of the vehicle, a workshop visit will be proposed. Due to the transmitted information I required spare parts can be delivered in advance to the workshop.

If the chassis of the vehicle designed as an adaptive chassis, a chassis setting of the chassis can be changed so that the actual suspension behavior is adapted to the learned optimal suspension behavior. Ie. In particular, the control unit 4 be used to adjust the suspension so that learned during vehicle development, ie trained, optimal suspension behavior is restored or the actual suspension behavior is at least approximated to the learned optimal suspension behavior.

Advantageously, this form the adaptive chassis and the control unit 4 a control cycle 2 , As the adaptive suspension alters parameters of its settings, the controller analyzes and evaluates 4 the settings and provides a control signal RS For example, a quality of the current settings and / or an evaluation of the parameter change and / or an instruction or an indication of the direction in which the parameters are to be changed to the adaptive chassis. 1 shows a schematic representation of a device 1 to carry out this process V , wherein the chassis of the vehicle is designed here as an adaptive suspension. Therefore, the device includes 1 here the control cycle 2 to adapt this adaptive suspension.

By means of internal and / or external sensors 3 become input signals IT of the driver, for example steering movements, vehicle sensor data FS of the vehicle, for example vehicle accelerations, and context sensor data KS with respect to the vehicle context, for example, an outside temperature and / or the number of people in the vehicle detected, and to the control unit 4 which is at the same time an evaluation unit for this data or comprises such an evaluation unit. In the control unit 4 is in the process V to monitor the chassis checked whether a vehicle movement and vehicle control and thus the actual suspension behavior is consistent with the learned optimal suspension behavior.

If this is not the case, optimization takes place O the adaptive suspension. Here, the suspension settings with respect to a chassis wear and / or chassis damage, which was detected by means of the comparison between the learned optimal suspension behavior and the actual suspension behavior optimized.

The control unit 4 gives the corresponding control signal RS to a chassis control unit 5 of the adaptive suspension. This control signal RS includes, for example, the instruction that the suspension settings should be optimized and / or which suspension parameters were or should be changed. From the chassis control unit 5 of the adaptive suspension, the controller 4 a feedback signal FBS , ie a feedback signal, received.

A modified in this way suspension behavior gF is by means of internal and / or external sensors 3 detected.

In addition to this adaptation of the adaptive suspension, the controller gives 4 an information signal IS with the corresponding information I in that the actual suspension behavior deviates from the learned optimum suspension behavior and / or that there is chassis wear and / or chassis damage to the output unit 6 , In particular head unit, the vehicle from which this information I outputs to the driver and / or outputs, for example, a proposal for maintenance of the landing gear.
Alternatively or additionally, an information signal IS with the corresponding information I in that the actual suspension behavior deviates from the learned optimum suspension behavior and / or that there is chassis wear and / or chassis damage to the vehicle-external device 7 , in particular of the manufacturer of the vehicle, in particular to the data center of the manufacturer of the vehicle issued, from which the vehicle manufacturer information I relating to the chassis wear and / or chassis damage of his vehicles in the field, ie in normal driving at customers relate.

2 schematically shows a flowchart for learning the optimal suspension behavior, which is the basis of this method V for monitoring the chassis of the vehicle.

After a training start TS , also referred to as learning start, takes place in a first training step TS1 a detection of steering movements and other sensor data of the vehicle internally and / or externally.

Subsequently, in a second training step TS2 It has been checked whether the suspension behavior has already been learned sufficiently well, ie the method of machine learning has been trained, so that the control signals and the suspension behavior are consistent.

If so, in 2 marked with the reference numeral J for yes, this is the end of training TE Also known as learners, this training, ie the learning of the optimal suspension behavior achieved.

If not, in 2 marked with the reference numeral N for no, takes place in a third training step TS3 a further learning of the suspension behavior and then a return to the first training step TS1 ,

The respective training step TS1 . TS2 . TS3 can also be called a learning step.

This learning of the optimal suspension behavior can be, for example, online, d. H. during test drives, and / or offline, d. H. using collected data from test drives.

3 schematically shows a flowchart of the method V for monitoring the chassis of the vehicle.

After a surveillance start US takes place in a first monitoring step US1 a detection of steering movements and other sensor data of the vehicle internally and / or externally.

Subsequently, in a second monitoring step US2 checked whether control signals and thus the actual suspension behavior, especially over a longer period of time, deviate from the expected learned optimal suspension behavior.

If not, in 3 marked with the reference numeral N for no, that's the end of surveillance UE achieved this monitoring.

The control signals and thus the actual suspension behavior, especially over a longer period of time, deviate from the expected optimum chassis behavior 3 marked with the reference numeral J for yes, will be in a third monitoring step US3 a corresponding information I , in particular via an existing chassis wear and / or chassis damage by means of the output unit 6 , in particular the head unit, of the vehicle output and optionally in a fourth monitoring step US4 a corresponding information I , in particular via the existing chassis wear and / or chassis damage, as field data to the vehicle-external device 7 , in particular the manufacturer of the vehicle, in particular sent to the data center of the manufacturer of the vehicle.

Subsequently, the monitoring end UE achieved this monitoring.

4 schematically shows a flowchart for optimization O Chassis settings of the adaptive chassis of the vehicle.

After an optimization start OS takes place in a first optimization step OS1 a detection of steering movements and other sensor data of the vehicle internally and / or externally.

Subsequently, in a second optimization step OS2 checked whether control signals and thus the actual suspension behavior, especially over a longer period of time, deviate from the expected learned optimal suspension behavior.

If not, in 4 marked with the reference numeral N for no, that's the end of optimization OE achieved this optimization.

The control signals and thus the actual suspension behavior, especially over a longer period of time, deviate from the expected optimum chassis behavior 4 marked with the reference numeral J for yes, will be in a third optimization step OS3 evaluated how, especially in which direction, the chassis has to be optimized.

In a fourth optimization step OS4 be from the controller 4 and / or the chassis control unit 5 Changed the suspension settings and then returned to the first optimization step OS1 , The optimization steps OS1 to OS4 are thus part of the control cycle 2 ,

LIST OF REFERENCE NUMBERS

1

contraption

2

Control circuit

3

sensor

4

control unit

5

Suspension control unit

6

output unit

7

vehicle-external device

IT

input

FBS

Feedback signal

FS

Vehicle sensor data

gF

Changed chassis behavior

I

information

J

Yes

IS

information signal

KS

Context sensor data

N

No

O

Optimize

OE

optimizing end

OS

optimization start

OS1

first optimization step

OS2

second optimization step

OS3

third optimization step

OS4

fourth optimization step

RS

control signal

TE

end of training

TS

training start

TS1

first training step

TS2

second training step

TS3

third training step

UE

monitoring end

US

monitoring start

US1

first monitoring step

US2

second monitoring step

US3

third monitoring step

US4

fourth monitoring step

V

method

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• DE 102009001689 A1 [0002]

Claims (3)

Process (V) for monitoring a chassis of a vehicle, characterized in that an actual suspension behavior is compared with a learned by machine learning optimum suspension behavior and will be closed at a deviation of the actual suspension behavior from the learned optimal suspension behavior to a chassis wear and / or landing gear damage, wherein then - a corresponding information (I) is output by means of at least one output unit (6) of the vehicle, and / or - a corresponding information (I) is transmitted to an off-vehicle device (7), and / or - a suspension setting of the as an adaptive Chassis trained chassis is changed so that the actual suspension behavior is adapted to the learned optimal suspension behavior. Method (V) according to Claim 1 , characterized in that the corresponding information (I) is output by means of a trained as a head unit output unit (6) of the vehicle. Method (V) according to one of the preceding claims, characterized in that the corresponding information (I) is transmitted to a vehicle-external device (7) of a manufacturer of the vehicle, in particular to a data center of the manufacturer of the vehicle.

Images