Detection and reconstruction of sensor faults

• US 10,204,461 B2
• Filed: 07/19/2016
• Issued: 02/12/2019
• Est. Priority Date: 07/19/2016
• Status: Active Grant

• Current Assignee: GM Global Technology Operations LLC (General Motors Company), University of Waikato
• Original Assignee: GM Global Technology Operations LLC (General Motors Company), University of Waikato
• Inventors: Chen, Shih-Ken, Moshchuk, Nikolai K., Litkouhi, Bakhtiar B., Zarringhalam, Reza, Khajepour, Amir, Melek, William
• Primary Examiner(s): Odeh, Nadeem
• Application Number:
  US15/214,162
• Publication Number:
  US 20180025558A1
• Time in Patent Office: 938 Days
• Field of Search: 701 297, 701 302, 701 307
• CPC Class Codes: B60W 2050/0025 Transfer function weighting factor
• CPC Class Codes: B60W 2050/0215 Sensor drifts or sensor failures
• CPC Class Codes: B60W 2050/0292 Fail-safe or redundant systems, e.g. limp-home or backup systems
• CPC Class Codes: B60W 2520/105 Longitudinal acceleration
• CPC Class Codes: B60W 2520/125 Lateral acceleration
• CPC Class Codes: B60W 2520/14 Yaw
• CPC Class Codes: B60W 2540/18 Steering angle
• CPC Class Codes: B60W 50/00 Details of control systems for road vehicle drive control not related to the control of a particular sub-unit , e.g. process diagnostic or vehicle driver interfaces
• CPC Class Codes: B60W 50/0205 Diagnosing or detecting failures; Failure detection models
• CPC Class Codes: B60W 50/029 Adapting to failures or work around with other constraints, e.g. circumvention by avoiding use of failed parts
• CPC Class Codes: G05B 2219/37537 Virtual sensor
• CPC Class Codes: G05B 23/0221 Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
• CPC Class Codes: G07C 5/0808 Diagnosing performance data testing of vehicles G01M17/00; testing of electrical installation on vehicles G01R31/005

Abstract

Methods and systems are provided for detecting faults in a sensor and reconstructing an output signal without use of the faulty sensor. In one embodiment, a method includes: receiving, by a processor, sensor data indicating a measured value from a first sensor; receiving, by a processor, sensor data indicating measured values from a plurality of other sensors; computing, by a processor, virtual values based on a vehicle model and the sensor data from the plurality of other sensors; computing, by a processor, a residual difference between the measured value from the first sensor and the virtual values; detecting, by a processor, whether a fault exists in the first sensor based on the residual difference; and when a fault in the sensor is detected, generating, by a processor, a control value based on the virtual values instead of the measured value.

Claims

1. A computer-implemented method, comprising:

receiving, by a processor, sensor data indicating a measured value from a first sensor;

receiving, by the processor, sensor data indicating measured values from a plurality of other sensors;

computing, by the processor, a first virtual value of the first sensor based on a first vehicle model and the sensor data from a first subset of the plurality of other sensors;

computing, by the processor, a second virtual value of the first sensor based on a second vehicle model and the sensor data from a second subset of the plurality of other sensors;

computing, by the processor, a first residual difference between the measured value from the first sensor and the first virtual value;

computing, by the processor, a second residual difference between the measured value from the first sensor and the second virtual value;

detecting, by the processor, that a fault exists in the first sensor based on the first residual difference, the second residual difference and a dynamic adaptive fault threshold;

when the fault in the first sensor is detected, reconstructing, by the processor, a sensor signal based on the first virtual value, the second virtual value, and computed weights applied to the first virtual value and the second virtual value; and

generating, by the processor, a control value based on the sensor signal, instead of the measured value of the first sensor, and for use in control of a vehicle,

wherein the dynamic adaptive fault threshold is a larger one of a current instantaneous adaptive fault threshold and an averaged instantaneous adaptive fault threshold, wherein the averaged instantaneous adaptive fault threshold occurs over a dynamic threshold window size, and wherein the window size is representative of a window time and is set by a predetermined number of previous instantaneous adaptive fault thresholds.

2. The method of claim 1, wherein the detecting that the fault exists comprises comparing the first and second residual differences with the dynamic adaptive fault threshold.

3. The method of claim 2, wherein the comparing the first and second residual differences with the dynamic adaptive fault threshold comprises detecting that the first and second residual differences exceed the dynamic adaptive fault threshold for a predetermined time period.

4. The method of claim 3, wherein the dynamic adaptive fault threshold is computed using at least one of a driving condition and a dynamic region.

5. The method of claim 1, wherein the reconstructing the sensor signal is based on a weighted average of the first virtual value and the second virtual value.

6. The method of claim 1, wherein the first sensor is a lateral acceleration sensor, and wherein the plurality of other sensors includes a steering angle sensor, an angular velocity sensor, and a yaw rate sensor.

7. The method of claim 1, wherein the first sensor is a yaw rate sensor, and wherein the plurality of other sensors includes a steering angle sensor, an angular velocity sensor, a lateral acceleration sensor, and a longitudinal velocity sensor.

8. The method of claim 1, wherein the first sensor is a longitudinal acceleration sensor, and wherein the plurality of other sensors includes a steering angle sensor, an angular velocity sensor, and a longitudinal velocity sensor.

9. A system, comprising:

a first non-transitory module that receives, by a processor, sensor data indicating a measured value from a first sensor, that receives, by the processor, sensor data indicating measured values from a plurality of other sensors, that computes, by the processor, a first virtual value of the first sensor based on a first vehicle model and a first subset of the sensor data from the plurality of other sensors, and that computes, by the processor, a second virtual value of the first sensor based on a second vehicle model and the sensor data from a second subset of the plurality of other sensors;

a second non-transitory module that computes, by the processor, a first residual difference between the measured value from the first sensor and the first virtual value, that computes, by the processor, a second residual difference between the measured value from the first sensor and the second virtual value; and

a third non-transitory module that detects, by the processor, that a fault exists in the first sensor based on the first residual difference, the second residual difference and a dynamic adaptive fault threshold, and when the fault in the first sensor is detected, reconstructs, by the processor, a sensor signal based on the first virtual value, the second virtual value, and computed weights applied to the first virtual value and the second virtual value, and generates, by the processor, a control value based on the sensor signal instead of the measured value of the first sensor, and for use in control of a vehicle,

wherein the dynamic adaptive fault threshold is a larger one of a current instantaneous adaptive fault threshold and an averaged instantaneous adaptive fault threshold, wherein the averaged instantaneous adaptive fault threshold occurs over a dynamic threshold window size, and wherein the window size is representative of a window time and is set by a predetermined number of previous instantaneous adaptive fault thresholds.

10. The system of claim 9, wherein the third non-transitory module detects that the fault exists by comparing the first and second residual differences with the dynamic adaptive fault threshold.

11. The system of claim 10, wherein the third non-transitory module compares the first and second residual differences with the dynamic adaptive fault threshold by detecting that the first and second residual differences exceed the dynamic adaptive fault threshold for a predetermined time period.

12. The system of claim 9, wherein the dynamic adaptive fault threshold is computed using at least one of a driving condition and a dynamic region.

13. The system of claim 9, wherein the third non-transitory module reconstructs the sensor signal based on a weighted average of the first virtual value and the second virtual value.

14. The system of claim 9, wherein the first sensor is a lateral acceleration sensor, and wherein the plurality of other sensors includes a steering angle sensor, an angular velocity sensor, and a yaw rate sensor.

15. The system of claim 9, wherein the first sensor is a yaw rate sensor, and wherein the plurality of other sensors includes a steering angle sensor, an angular velocity sensor, a lateral acceleration sensor, and a longitudinal velocity sensor.

16. The system of claim 9, wherein the first sensor is a longitudinal acceleration sensor, and wherein the plurality of other sensors includes a steering angle sensor, an angular velocity sensor, and a longitudinal velocity sensor.