Method and system for predicting system status

1. A method of determining status of a test specimen associated with a vehicle system, the method comprising:

collecting, by a computing system, a plurality of data associated with a test specimen and the vehicle system;

selecting a relevant data set of the plurality of data, comprising an iterative process of identification and extraction of representative variables from the plurality of data, the extraction based on a plurality of calculated correlation coefficients between each pair of variables of the plurality of data, wherein the extraction includes:

determining a correlation coefficient for each pair of variables in the plurality of data based on an average correlation coefficient for each pair of variables in the plurality of data over a selected number of operating cycles;

computing a sum square of the correlation coefficients;

ascertaining a set of variables exhibiting the highest correlation, yet not duplicative with correlation coefficients for other variables; and

extracting only variables that are not duplicative;

transforming at least a portion of the selected relevant data set to form a transformed data set, the transforming based on mathematical properties;

collecting statistics associated with the selected relevant data set and the transformed data set to form a statistics data set;

classifying the selected relevant data set, transformed data set and the statistics data set; and

predicting the status of the test specimen based on the classifying.

2. The method of claim 1 , wherein the collecting comprises operating the vehicle system and acquiring the plurality of data over at least a portion of an operating cycle.

3. The method of claim 1 , wherein the plurality of data comprises sensor data, an operating condition of the vehicle or a sensor, or status of any of the vehicle system, a sensor system, or the sensor.

4. The method of claim 1 , wherein the plurality of data comprises a time history.

5. The method of claim 1 , wherein the test specimen includes a sensor system and the sensor system includes an oxygen sensor, the oxygen sensor disposed in an exhaust gas stream of a motor vehicle.

6. The method of claim 1 , wherein the transforming includes at least one of averaging, spectral filtering, derivatives, integrals, trigonometric functions, linear and non-linear mathematical functions, and statistical functions.

7. The method of claim 6 , wherein the transforming is a derivative over a selected interval.

8. The method of claim 7 , wherein the selected interval is a single sample interval.

9. The method of claim 1 , wherein the collecting statistics comprises computing at least one of a percentile, a mean, and a standard deviation for at least one piece of data in the selected relevant data set and the transformed data set to form the statistics data set.

10. The method of claim 1 , wherein the classifying the selected relevant data set, transformed data set, and the statistics data set comprises applying the selected relevant data set, transformed data set, and the statistics data set to a neural network.

11. The method of claim 10 , wherein the neural network comprises an input layer having 308 inputs a first hidden layer have 400 nodes, and an output layer having 3 nodes, the three nodes of the output corresponding to a class associated with a condition of the sensor.

12. The method of claim 10 , further comprising training the neural network based on iterating through the at least a portion of the selected relevant data set, transformed data set, and the statistics data set to adjust weights given to each neuron of the neural network.

13. The method of claim 1 , wherein the predicting a status of the test specimen based on the classifying identifies the test specimen exhibiting at least one of three states, the three states defined as an OK state, a Pre-alarm state, and an Alarm state.

14. A system for determining a status of a test specimen, the system comprising:

a test specimen associated with a vehicle system having at least a sensor system, the sensor system having at least one sensor; and

a computing system in operable communication with at least the at least one sensor of the sensor system and the vehicle system, the computing system configured to execute a method of determining status of the test specimen comprising:

collecting, by the computing system, a plurality of data associated with a test specimen;

selecting a relevant data set of the plurality of data, comprising an iterative process of identification and extraction of representative variables from the plurality of data, the extraction based on a plurality of calculated correlation coefficients between each pair of variables of the plurality of data, wherein the extraction includes:

determining a correlation coefficient for each pair of variables in the plurality of data based on an average correlation coefficient for each pair of variables in the plurality of data over a selected number of operating cycles;

computing a sum square of the correlation coefficients;

ascertaining a set of variables exhibiting the highest correlation, yet not duplicative with correlation coefficients for other variables; and

extracting only variables that are not duplicative;

transforming at least a portion of the selected relevant data set to form a transformed data set, the transforming based on mathematical properties;

collecting statistics associated with the selected relevant data set and the transformed data set to form a statistics data set;

classifying the selected relevant data set, transformed data set and the statistics data set; and

predicting the status of the test specimen based on the classifying.

15. The system of claim 14 , wherein the plurality of data comprises at least one of sensor data, an operating condition of the vehicle or the sensor, and status of any of the vehicle system, a sensor system and the sensor.

16. The system of claim 14 , wherein the sensor system includes an oxygen sensor, the oxygen sensor disposed in an exhaust gas stream of a motor vehicle.