Method and system for detecting and localizing sensor defects in motor vehicles

• US 6,470,300 B1
• Filed: 04/28/1999
• Issued: 10/22/2002
• Est. Priority Date: 04/28/1998
• Status: Expired due to Term

• Current Assignee: 21ST CENTURY GARAGE LLC
• Original Assignee: Daimler Chrysler Company LLC (Stellantis NV)
• Inventors: Mahmoud, Rachad, Benzinger, Matthias, Suissa, Avshalom, Boettiger, Friedrich
• Primary Examiner(s): WACHSMAN, HAL D
• Application Number:
  US09/300,590
• Time in Patent Office: 1,273 Days
• Field of Search: 702/190, 702/35, 702/36, 702/38, 702/57, 702/96, 702/104, 702/108, 702/105, 702/113-118, 702/124, 702/126, 702/127, 702/141, 702/142, 702/145-148, 702/154, 702/182-185, 702/188, 702/189, 702/193, 702/194, 702/196, 702/197, 702/FOR.160, 702/FOR.164, 702/FOR.103, 702/FOR-170, 702/FOR-123, 702/FOR.104, 702/FOR.156, 702/FOR.157, 702/FOR.134, 702/FOR.135, 702/FOR.141, 702/FOR.150, 702/FOR.151, 702/FOR.155, 701/29-31, 701/34, 701/35, 701/58-61, 701/63, 701/65, 701/70, 701/71, 701/74, 701/79, 701/82-84, 701/87, 701/90-95, 703/2, 703/6-8, 731/16, 731/172, 731/173, 731/21, 700/29-31, 303/150
• US Class Current: 702/190
• CPC Class Codes: G05B 9/02 electric

Abstract

For detecting and localizing sensor defects in motor vehicles, a sensor determines a measuring signal for the description of the dynamic behavior of the motor vehicle. A state value assigned to the measuring signal is computed from a mathematical equivalent model. An estimation error of the state value is determined and a residue is determined from the difference between the measuring signal and a reference quantity which corresponds to the measuring signal and is formed from the state value. For clearly detecting and localizing the sensor defects, a characteristic parameter is formed by multiplying the residue and the estimation error, and an error signal is generated if the characteristic parameter exceeds a limit value.

Claims

1. A method for detecting and localizing sensor defects in a motor vehicle, the method comprising the acts of:

determining a measuring signal (y_k) for a description of a dynamic behavior of the motor vehicle by a sensor;

computing from a mathematical equivalent model, a state value (x⁻

_k) which is assigned to the measuring signal (y_k);

computing an estimation error (e⁻

_k) of the state value (x⁻

_k);

determining a residue (r_k) from the difference between the measuring signal (y_k) and a reference quantity (C_k*x⁻

_k) corresponding to the measuring signal (y_k) and formed from the state value (x⁻

_k);

forming a characteristic parameter (ε

) by which sensor defects are detected and localized from the multiplication of the residue (r_k) and the estimation error (e⁻

_k); and

generating an error signal (E_I, E_L) when the characteristic parameter (ε

) exceeds a limit value (s_k) to indicate that a defect in at least one of the pertaining sensors was detected (E_I) and to indicate which one of the pertaining sensors is defective (E_L).

2. The method according to claim 1, wherein the mathematical equivalent model is based on a Kalman filter algorithm, the state value being identical to a prediction value and corresponding to the equation

3. The method according to claim 2, wherein the residue (r_k) is computed from the measuring signal (y_k) and the state value (x⁻

_k) according to the equation

4. The method according to claim 2, wherein an optimal filter estimation value (x⁺_k) is determined from the rule

5. The method according to claim 1, wherein the residue (r_k) is computed from the measuring signal (y_k) and the state value (x⁻

_k) according to the equation

6. The method according to claim 5, wherein an optimal filter estimation value (x⁺_k) is determined from the rule

7. The method according to claim 1, wherein an optimal filter estimation value (x⁺_k) is determined from the rule

8. The method according to claim 7, wherein the amplification matrix (K_k) is a Kalman gain which is computed according to the equation

9. The method according to claim 8, wherein the estimation error (e⁻

_k) is determined from the difference between the optimal filter estimation value (x⁺<sub>k−

1</sub>) and the prediction value (x⁻

_k) according to the equation

10. The method according to claim 9, wherein the characteristic parameter (ε

) is determined from the equation

11. The method according to claim 7, wherein the estimation error (e⁻

_k) is determined from the difference between the optimal filter estimation value (x⁺<sub>k−

1</sub>) and the prediction value (x⁻

_k) according to the equation

12. The method according to claim 11, wherein the characteristic parameter (ε

) is determined from the equation

13. The method according to claim 1, wherein the limit value (s_k) is adaptively calculated from the residue (r_k) according to the rule

14. The method according to claim 1, wherein for several sensors the vector of measuring signals (y_k) takes the form

15. The method according to claim 14, wherein a weighting matrix (M_k) has the form $M_k=\left[\begin{array}{ccccc}{m}_{11}& 0& m_{13}& 0& 0\\ 0& m_{22}& 0& m_{24}& 0\\ m_{31}& 0& m_{33}& 0& 0\\ 0& m_{42}& 0& m_{44}& 0\\ 0& 0& 0& 0& m_{55}\end{array}\right]$

wherein the coefficients (M_ij) of the weighting matrix (M_k) are determined empirically.

16. The method according to claim 15, wherein the vector of state values (x⁻

_k) has the form

17. The method according to claim 15, wherein the diagonal matrix (R^M_k), which contains the components of the residue (r_k) on the main diagonal, has the form $R_k^M=[\begin{array}{c}{r}_{1}\end{array}$

k00000r2

k00000r3

k00000r4

k00000r5

k]

the estimation error (e⁻

_k) only taking into account the differences of inherently dynamic state quantities.

18. The method according to claim 15, wherein an input vector (u_k) has the form

19. The method according to claim 14, wherein the vector of state values (x⁻

_k) has the form

20. The method according to claim 19, wherein the diagonal matrix (R^M_k), which contains the components of the residue (r_k) on the main diagonal, has the form $R_k^M=[\begin{array}{c}{r}_{1}\end{array}$

k00000r2

k00000r3

k00000r4

k00000r5

k]

the estimation error (e⁻

_k) only taking into account the differences of inherently dynamic state quantities.

21. The method according to claim 19, wherein an input vector (u_k) has the form

22. The method according to claim 14, wherein the diagonal matrix (R^M_k), which contains the components of the residue (r_k) on the main diagonal, has the form $R_k^M=[\begin{array}{c}{r}_{1}\end{array}$

k00000r2

k00000r3

k00000r4

k00000r5

k]

the estimation error (e⁻

_k) only taking into account the differences of inherently dynamic state quantities.

23. The method according to claim 22, wherein an input vector (u_k) has the form

24. The method according to claim 14, wherein an input vector (u_k) has the form

25. A system for detecting and localizing sensor defects in a motor vehicle, comprising:

sensors which feed measuring signals (y_k) describing a dynamic behavior of the motor vehicle as input signals;

a computing unit which receives the measuring signals;

an evaluation unit coupled to an output side of the computing unit;

a computer readable medium for the computing unit and the evaluation unit, said computer readable medium having stored thereon program code segments that;

use a mathematical equivalent model in computing state values (x⁻

_k) assigned to the measuring signals;

compute estimation errors (e⁻

_k) of the state values;

determine residues (r_k) from differences between the measuring signals and reference quantities corresponding to the measuring signals and formed from the state values;

form characteristic parameters (ε

) via multiplication of the residues and the estimation errors; and

generate error signals (E_I, E_L) when the characteristic parameters exceed limit values (s_k) to indicate defects in pertaining sensors were detected (E_I) and to indicate which of the pertaining sensors were defective (E_L).