Vehicle data analysis apparatus, vehicle data analysis method, and defect diagnosis apparatus

• US 9,177,425 B2
• Filed: 03/19/2012
• Issued: 11/03/2015
• Est. Priority Date: 03/22/2011
• Status: Active Grant

• Current Assignee: Toyota Jidosha Kabushiki Kaisha (Toyota Motor Corporation)
• Original Assignee: Toyota Jidosha Kabushiki Kaisha (Toyota Motor Corporation)
• Inventors: Suganuma, Hideaki, Matsumura, Takeshi, Hikita, Toshiro
• Primary Examiner(s): Nguyen, John Q
• Assistant Examiner(s): KERRIGAN, MICHAEL V
• Application Number:
  US14/005,838
• Publication Number:
  US 20140095016A1
• Time in Patent Office: 1,324 Days
• US Class Current: 1/1
• 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: G06F 15/00 Digital computers in general details G06F1/00 – G06F13/00; Data processing equipment in general
• CPC Class Codes: G07C 5/00 Registering or indicating the working of vehicles

Abstract

A vehicle data analysis apparatus analyzes vehicle data that indicates chronological change of a vehicle state. The vehicle data analysis apparatus is provided with a computing unit and a recognition unit. The computing unit is configured to make obvious data fluctuation accompanying the development, in the vehicle data, of mechanical or control-related fault in a vehicle control system. The recognition unit is configured to recognize the vehicle data to be considered during vehicle fault diagnosis on the basis of a result of computation by the computing unit.

Claims

1. A vehicle data analysis apparatus for analyzing vehicle data indicating time-series changes of a state of a vehicle, the apparatus comprising:

a storage device in which a plurality of the vehicle data obtained from the vehicle are accumulated as data to be analyzed;

a computing section configured to uncover a data fluctuation in each of the plurality of the vehicle data accumulated in the storage device that results from a mechanical or control-related abnormality in a vehicle control system and calculate an amount of change per unit time of the vehicle data; and

a recognizing section configured to recognize, based on a computation result obtained by the computing section, vehicle data to be subject to a diagnosis of the abnormality in the vehicle as a diagnosis target by excluding the data including the amount of change smaller than a threshold showing the data fluctuation resulting from a driver'"'"'s operation of the vehicle from the data fluctuation included in the vehicle data,

wherein the recognizing section is configured to use the threshold set to exclude the data fluctuation resulting from a human factor.

2. The vehicle data analysis apparatus according to claim 1, wherein

the vehicle data includes, as data reflecting a characteristic of a driver'"'"'s operation of the vehicle, a sensor dataset and a control dataset, the sensor dataset indicating changes of a value detected by a sensor that detects the state of the vehicle, and the control dataset reflecting changes of the sensor dataset, and

the recognizing section is configured to, based on the computation result obtained by the computing section, identify, as the diagnosis target, vehicle data that includes a data fluctuation based on a factor different from the driver'"'"'s vehicle operation of the vehicle.

3. The vehicle data analysis apparatus according to claim 1, wherein

the computing section also performs a computation to identify vehicle data that indicates that the calculated amount of change is equal to or greater than a threshold value that indicates an amount of change caused by data fluctuation resulting from mechanical or control-related abnormality in the vehicle control system.

4. A fault diagnosis apparatus, comprising:

the vehicle data analysis apparatus according to claim 1,

wherein the fault diagnosis apparatus is configured to diagnose the presence or absence of any fault in the vehicle based on the vehicle data analyzed by the vehicle data analysis apparatus.

5. A vehicle data analysis apparatus for analyzing vehicle data indicating time-series changes of a state of a vehicle, the apparatus comprising:

a storage device in which a plurality of the vehicle data obtained from the vehicle are accumulated as data to be analyzed;

a computing section configured to uncover a data fluctuation in each of the plurality of the vehicle data accumulated in the storage device that results from a mechanical or control-related abnormality in a vehicle control system, wherein as the computation, the computing section performs a computation for obtaining statistical distribution of the vehicle data stored in the storage device; and

a recognizing section configured to;

recognize, based on a computation result obtained by the computing section, vehicle data to be subject to a diagnosis of the abnormality in the vehicle as a diagnosis target by excluding vehicle data deviating from a median in the obtained statistical distribution showing the data fluctuation resulting from a driver'"'"'s operation of the vehicle from the data fluctuation included in the vehicle data, and

extract vehicle data that has a relatively large deviation in the recognized vehicle data from the storage device as the diagnosis target vehicle data,

wherein the recognizing section is configured to use a threshold set to exclude the data fluctuation from a human factor in excluding vehicle data deviating from the median.

6. The vehicle data analysis apparatus according to claim 5, wherein

the vehicle data includes a sensor dataset that indicates changes of a value detected by a sensor, which detects the state of the vehicle, and a control dataset reflecting, as data corresponding to the sensor dataset, changes of the corresponding sensor dataset, and

the sensor dataset and the control dataset are associated with each other as a result of determination that they are highly correlated with each other on condition that:

a. time-series changes correlate with each other; and

b. delay time of the control dataset with respect to the sensor dataset is within the response time of a vehicle'"'"'s behavior with respect to the driver'"'"'s operation of the vehicle, the response time being indicated by the sensor dataset.

7. The vehicle data analysis apparatus according to claim 5, further comprising a correspondence table representing a correspondence between vehicle data of a plurality of types,

wherein, in the correspondence table, a sensor dataset indicating changes of a value detected by a sensor that detects the state of the vehicle and a control dataset highly correlated with the sensor dataset are related to each other.

8. The vehicle data analysis apparatus according to claim 7, wherein

in the correspondence table, as a sensor dataset and a control dataset highly correlated with each other, a dataset indicating changes of a value detected from an accelerator pedal of the vehicle and a dataset indicating changes of the amount of control of an engine control device for controlling an engine installed in the vehicle are registered in relation to each other, and

a dataset indicating changes of the amount of operation of a steering wheel of the vehicle and a dataset indicating changes of the amount of control of a VSC control device for exerting stability control for the vehicle are also registered in relation to each other.

9. A vehicle data analysis apparatus for analyzing vehicle data indicating time-series changes of a state of a vehicle, the apparatus comprising:

a storage device in which a plurality of the vehicle data obtained from the vehicle are accumulated as data to be analyzed, the vehicle data including a sensor dataset and a control dataset;

a computing section configured to uncover a data fluctuation in each of the plurality of the vehicle data accumulated in the storage device that results from a mechanical or control-related abnormality in a vehicle control system and obtain, in the computation, a difference between the sensor dataset and the control dataset correlated with the sensor dataset; and

a recognizing section configured to;

recognize, based on a computation result obtained by the computing section, vehicle data to be subject to a diagnosis of the abnormality in the vehicle as a diagnosis target by excluding a sensor dataset and a control dataset whose difference is not equal to or greater than a threshold showing the data fluctuation resulting from a driver'"'"'s operation of the vehicle from the data fluctuation included in the vehicle data, and

recognize the diagnosis target vehicle data based on the obtained difference,

wherein the recognizing section is configured to use the threshold set to exclude the data fluctuation resulting from a human factor.

10. The vehicle data analysis apparatus according to claim 9, wherein the computing section calculates the difference between the sensor dataset and the control dataset corresponding to the sensor dataset based on the sensor dataset and the control dataset that have been subjected to frequency analysis.

11. The vehicle data analysis apparatus according to claim 9, wherein

the sensor dataset and the control dataset are associated with each other as a result of determination that they are highly correlated with each other on condition that:

a. time-series changes correlate with each other; and

b. delay time of the control dataset with respect to the sensor dataset is within the response time of a vehicle'"'"'s behavior with respect to the driver'"'"'s operation of the vehicle, the response time being indicated by the sensor dataset.

12. A vehicle data analysis method for analyzing vehicle data indicating time-series changes of a state of a vehicle using a vehicle data analysis apparatus, the method comprising:

accumulating a plurality of the vehicle data obtained from the vehicle in a storage device as data to be analyzed;

uncovering a data fluctuation in each of the plurality of the vehicle data that results from a mechanical or control-related abnormality in a vehicle control system;

calculating an amount of change per unit time of the vehicle data;

identifying the vehicle data as a diagnosis target to diagnose the abnormality in the vehicle by excluding the data including the amount of change smaller than a threshold showing the data fluctuation resulting from a driver'"'"'s operation of the vehicle from the data fluctuation included in the vehicle data; and

using the threshold set to exclude the data fluctuation resulting from a human factor.

13. The vehicle data analysis method according to claim 12, wherein

the identification is to select vehicle data that indicates that the calculated amount of change per unit time in the vehicle data is equal to or greater than a threshold value that indicates an amount of change caused by data fluctuation resulting from the mechanical or control-related abnormality in the vehicle control system.

14. A vehicle data analysis method for analyzing vehicle data indicating time-series changes of a state of a vehicle using a vehicle data analysis apparatus, the method comprising:

accumulating a plurality of the vehicle data obtained from the vehicle in a storage device as data to be analyzed, wherein the vehicle data includes, as data reflecting a characteristic of a driver'"'"'s operation of the vehicle, a sensor dataset and a control dataset, the sensor dataset indicating changes of a value detected by a sensor that detects the state of the vehicle, and the control dataset reflecting changes of the sensor dataset;

uncovering a data fluctuation in each of the plurality of the vehicle data that results from a mechanical or control-related abnormality in a vehicle control system; and

identifying the vehicle data as a diagnosis target to diagnose the abnormality in the vehicle by excluding the data fluctuation resulting from a driver'"'"'s operation of the vehicle from the data fluctuation included in the vehicle data, wherein

the identifying includes;

excluding a sensor dataset and a control dataset whose difference is not equal to or greater than a threshold, and identifying, as the diagnosis target, vehicle data that includes a data fluctuation based on a factor different from the driver'"'"'s vehicle operation of the vehicle, and

using the threshold set to exclude the data fluctuation resulting from a human factor.

15. The vehicle data analysis method according to claim 14, wherein the uncovering is to uncover a data fluctuation resulting from the mechanical or controls-related abnormality in the vehicle control system based on a difference between the sensor dataset and the control dataset correlated with the sensor dataset.

16. The vehicle data analysis method according to claim 15, wherein the uncovering is to subject the sensor dataset and the control dataset corresponding to the sensor dataset to frequency analysis and to calculate the difference between the sensor dataset and the control dataset based on the sensor dataset and the control dataset that have been subjected to the frequency analysis.

17. A vehicle data analysis method for analyzing vehicle data indicating time-series changes of a state of a vehicle using a vehicle data analysis apparatus, the method comprising:

accumulating a plurality of the vehicle data obtained from the vehicle in a storage device as data to be analyzed;

uncovering a data fluctuation in each of the plurality of the vehicle data that results from a mechanical or control-related abnormality in a vehicle control system, wherein the uncovering is to obtain statistical distribution of the stored vehicle data; and

identifying the vehicle data as a diagnosis target to diagnose the abnormality in the vehicle by excluding the data fluctuation resulting from a driver'"'"'s operation of the vehicle from the data fluctuation included in the vehicle data, wherein

the identifying includes;

excluding vehicle data deviating from a median in the vehicle data represented as a statistical distribution and identifying vehicle data that has a relatively large deviation among the recognized vehicle data as the diagnosis target vehicle data, and

using a threshold set to exclude the data fluctuation resulting from a human factor in excluding vehicle data deviating from the median.