Fault diagnostic method and device

• US 7,765,042 B2
• Filed: 06/03/2004
• Issued: 07/27/2010
• Est. Priority Date: 06/12/2003
• Status: Expired due to Fees

• Current Assignee: Robert Bosch GmbH
• Original Assignee: Robert Bosch GmbH
• Inventors: Weber, Reinhard, Foerstner, Dirk
• Primary Examiner(s): Hellner; Mark
• Assistant Examiner(s): Diacou; Ari M
• Application Number:
  US10/860,463
• Publication Number:
  US 20040254696A1
• Time in Patent Office: 2,245 Days
• Field of Search: 701 29- 35, 701/110, 701/115
• US Class Current: 701/33
• CPC Class Codes: B60T 2270/413 Plausibility monitoring, cross check, redundancy
• CPC Class Codes: B60T 8/885 using electrical circuitry
• CPC Class Codes: F02D 2041/1411 using a finite or infinite state machine, automaton or state graph for controlling or modelling
• CPC Class Codes: F02D 41/22 Safety or indicating devices for abnormal conditions in air/fuel ratio feedback systems F02D41/1495, in electric control linkage F02D11/107, in purge control systems F02M25/0809
• CPC Class Codes: G05B 23/0245 based on a qualitative model, e.g. rule based; if-then decisions
• CPC Class Codes: Y02T 10/40 Engine management systems

Abstract

A method for monitoring motor vehicle systems that include mechanical components and at least one control device detects, in a first step, input signals and output signals as quantitative signals that characterize the operating behavior, assigns discrete qualitative values to value intervals of the input signals and output signals, which are established using a particular sequence of threshold values. A change to one of the qualitative values is recognized as an event, and consecutive events are stored as an event sequence. The event sequence is compared with a model of the normal behavior of the motor vehicle system, in particular a model composed of all event sequences of the proper operation of the motor vehicle system. If there is a match, it is assumed that the motor vehicle system is operating properly. If there is no match, erroneous operation is assumed.

Claims

1. A method for monitoring a motor vehicle system having mechanical components and at least one control device, the method comprising:

detecting a plurality of input signals and a plurality of output signals as continuous-value signals that characterize the operating behavior of the motor vehicle system;

assigning discrete values to value intervals of the input signals and the output signals to form a plurality of discrete-value input signals and a plurality of discrete-value output signals, wherein the value intervals correspond to values between a plurality of pre-established thresholds, wherein at least one value interval is not equal in size to another value interval, wherein the discrete-value input signals are sequences of the discrete values that correspond to the values of the input signals, and wherein the discrete-value output signals are sequences of the discrete values that correspond to the values of the output signals;

combining the plurality of discrete-value input signals to form a single combined discrete-value input signal and combining the plurality of discrete-value output signals to form a single combined discrete-value output signal;

detecting as an event a change to at least one of;

the single combined discrete-value input signal or the single combined discrete-value output signal;

storing consecutive events as at least one signal event sequence;

comparing the at least one signal event sequence with at least one normal model comprising at least one normal event sequence, wherein the normal model is of a normal behavior of the motor vehicle system, and if the at least one signal event sequence matches the at least one normal event sequence then a determination is made that the motor vehicle system is operating properly, and if the at least one signal event sequence does not match the at least one normal event sequence then a determination is made that the motor vehicle system is operating erroneously.

2. The method of claim 1, wherein the at least one signal event sequence is compared with at least one fault model of the motor vehicle system, the at least one fault model comprising at least one fault event sequence which is associatable with a particular fault, and if there is a match, a presence of the particular fault is recognized.

3. The method of claim 2, wherein:

a presence of an unknown fault is recognized if at least one of;

the at least one signal event sequence matches the at least one fault event sequence of more than one fault model, or the at least one signal event sequence matches both the at least one normal event sequence of the normal model and the at least one fault event sequence of at least one fault model; and

a presence of an indeterminate fault is recognized if the at least one signal event sequence does not match any of;

the at least one normal event sequence of the normal model, and the at least one fault event sequence of any of the at least one fault models.

4. The method of claim 1, wherein at least two consecutive results of the comparing are compared with each other and this comparison is taken into consideration in determining whether a fault is present.

5. The method of claim 1, further comprising:

recording a time that has elapsed since an event occurred as a continuous-value time output signal, the continuous-value time output signal being one of the plurality of output signals;

assigning discrete values to value intervals of the time output signal to form discrete-value time output signal, wherein the discrete values are established using a particular sequence of threshold values, up to a maximum value, corresponding to the time output signal; and

wherein the combining the plurality of discrete-value output signals to form a single combined discrete-value output signal includes combining the discrete-value time output signal as part of the plurality of discrete-value output signals.

6. The method of claim 1, wherein the single combined discrete-value input signal is uniquely assignable to the original values of the plurality of discrete-value input signals, and the single combined discrete-value output signal is uniquely assignable to the original values of the plurality of discrete-value output signals.

7. The method of claim 1, wherein the plurality of discrete-value input and output signals are respectively combined as weighted sums.

8. The method of claim 1, wherein the single combined discrete-value input signal and the single combined discrete-value output signal are combined to form a single combined discrete value input-output signal.

9. The method of claim 1, wherein, in the assigning, the threshold values of a previous value interval are reduced by a lower hysteresis value and increased by an upper hysteresis value.

12. The method of claim 1, wherein the normal model is a complete normal model comprising all normal event sequences of a proper operation of the motor vehicle system.

13. The method of claim 1, wherein the at least one fault model is not a complete fault model, and does not include all fault event sequences of the motor vehicle system.

14. The method of claim 1, wherein the detecting detects as an event a change to the single combined discrete-value output signal.

15. The method of claim 8, wherein a previous value of the single combined discrete-value input signal, a previous value of the single combined discrete-value output signal, and a current value of the single combined discrete-value output signal are combined to form the single combined discrete value input-output signal.

16. The method of claim 1, further comprising:

recording a time that has elapsed since an event occurred as a continuous-value time output signal, the continuous-value time output signal being one of the plurality of output signals;

assigning discrete values to value intervals of the time output signal to form discrete-value time output signal, wherein the discrete values are established using a particular sequence of threshold values, up to a maximum value, corresponding to the time output signal;

wherein the combining the plurality of discrete-value output signals to form a single combined discrete-value output signal includes combining the discrete-value time output signal as part of the plurality of discrete-value output signals,

wherein the at least one signal event sequence is compared with at least one fault model of the motor vehicle system, the at least one fault model comprising at least one fault event sequence which is associatable with a particular fault, and if there is a match, a presence of the particular fault is recognized,

wherein a presence of an unknown fault is recognized if at least one of;

the at least one signal event sequence matches the at least one fault event sequence of more than one fault model, or the at least one signal event sequence matches both the at least one normal event sequence of the normal model and the at least one fault event sequence of at least one fault model,

wherein a presence of an indeterminate fault is recognized if the at least one signal event sequence does not match any of;

the at least one normal event sequence of the normal model, and the at least one fault event sequence of any of the at least one fault models, and

wherein, in the assigning, the threshold values of a previous value interval are reduced by a lower hysteresis value and increased by an upper hysteresis value.

17. The method of claim 1, wherein:

the single combined discrete-value input signal is uniquely assignable to the original values of the plurality of discrete-value input signals, and the single combined discrete-value output signal is uniquely assignable to the original values of the plurality of discrete-value output signals,

the plurality of discrete-value input and output signals are respectively combined as weighted sums, and

the single combined discrete-value input signal and the single combined discrete-value output signal are combined to form a single combined discrete value input-output signal.

18. The method of claim 1, wherein:

the normal model is a complete normal model comprising all normal event sequences of a proper operation of the motor vehicle system,

the at least one fault model is not a complete fault model, and does not include all fault event sequences of the motor vehicle system,

the detecting detects as an event a change to the single combined discrete-value output signal, and

a previous value of the single combined discrete-value input signal, a previous value of the single combined discrete-value output signal, and a current value of the single combined discrete-value output signal are combined to form the single combined discrete value input-output signal.

19. The method of claim 16, wherein:

the single combined discrete-value input signal is uniquely assignable to the original values of the plurality of discrete-value input signals, and the single combined discrete-value output signal is uniquely assignable to the original values of the plurality of discrete-value output signals,

the plurality of discrete-value input and output signals are respectively combined as weighted sums,

the single combined discrete-value input signal and the single combined discrete-value output signal are combined to form a single combined discrete value input-output signal,

the normal model is a complete normal model comprising all normal event sequences of a proper operation of the motor vehicle system,

the at least one fault model is not a complete fault model, and does not include all fault event sequences of the motor vehicle system,

the detecting detects as an event a change to the single combined discrete-value output signal, and

a previous value of the single combined discrete-value input signal, a previous value of the single combined discrete-value output signal, and a current value of the single combined discrete-value output signal are combined to form the single combined discrete value input-output signal.

10. A monitoring device for a motor vehicle system, including mechanical components and at least one control device, the monitoring device comprising:

an monitoring arrangement to monitor the motor vehicle system by performing the following;

detecting input signals and output signals as quantitative signals that characterize the operating behavior of the motor vehicle system;

assigning discrete qualitative values to value intervals of the input signals and the output signals to form a plurality of discrete-value input signals and a plurality of discrete-value output signals, wherein the value intervals correspond to values between a plurality of pre-established thresholds, wherein at least one value interval is not equal in size to another value interval, wherein the discrete-value input signals are sequences of the discrete values that correspond to the values of the input signals, and wherein the discrete-value output signals are sequences of the discrete values that correspond to the values of the output signals;

detecting a change to at least one of the qualitative values as an event;

storing consecutive events as at least one event sequence;

comparing the at least one stored event sequence with at least one model, each of which is composed of at least one event sequence, wherein the model is of a normal behavior of the motor vehicle system, and if the at least one stored event sequence matches the model of the normal behavior, then it is assumed that the motor vehicle system is operating properly, and erroneous operation is assumed if they do not match.

11. The monitoring device of claim 10, wherein the motor vehicle system includes a fuel injection system for an internal combustion engine, a braking system, or a by-wire system.