Method and apparatus for estimating state of battery

1. A battery state estimation apparatus comprising:

a sensing data acquirer configured to acquire sensing data on a battery; and

a battery state estimator configured to:

extract segments from the sensing data;

generate a segment vector by merging the segments;

obtain a feature vector by mapping the segment vector to a battery state space, wherein the battery state space comprises an area corresponding to each of battery state types; and

estimate a state of the battery based on a distribution of the feature vector in the battery state space.

2. The apparatus of claim 1 , wherein the battery state estimator comprises a sensing data segmenter configured to segment the sensing data based on a predetermined time interval, extract a segment having a predetermined size from each time interval.

3. The apparatus of claim 1 ,

wherein the battery state estimator further comprises a calibrator configured to correct a time error of the sensing data, using a determined calibration scheme, based on a predetermined calibration cycle as a reference, and

wherein the determined calibration scheme includes an interpolation scheme, a merging scheme, and an averaging scheme.

4. The apparatus of claim 1 , wherein the battery state estimator further comprises a feature space mapper configured to map the segment vector to the battery state space based on a predetermined mapping parameter.

5. The apparatus of claim 4 , wherein the predetermined mapping parameter comprises a predetermined reference matrix; and

the feature space mapper is further configured to project the segment vector onto the predetermined reference matrix to extract the feature vector having a dimension corresponding to the battery state space.

6. The apparatus of claim 5 , wherein the predetermined reference information comprises information on patterns for the battery state types in the battery state space; and

the battery state estimator further comprises a battery state determiner configured to determine the state of the battery by comparing the feature vector to the information on the patterns.

7. The apparatus of claim 6 , wherein the battery state types comprise a normal state type, an abnormal state type, and a fault state type; and

each of the abnormal state type and the fault state type comprises at least one subtype.

8. The apparatus of claim 6 , wherein the battery state determiner is further configured to calculate a similarity between the feature vector and each of the patterns in the battery state space, and determine the state of the battery based on a comparison between the calculated similarities.

9. A processor-implemented method of estimating a state of a battery, the method comprising:

acquiring sensing data on a battery;

extracting segments from the sensing data;

generating a segment vector by merging the segments;

obtaining a feature vector by mapping the segment vector to a battery state space, wherein the battery state space comprises an area corresponding to each of battery state types; and

estimating the state of the battery based on a distribution of the feature vector in the battery state space.

10. The method of claim 9 , wherein the extracting comprises:

segmenting the sensing data based on a predetermined time interval; and

extracting a segment having a predetermined size from each time interval.

11. The method of claim 9 ,

further comprises correcting a time error of the sensing data, using a determined calibration scheme, based on a predetermined calibration cycle as a reference, and

wherein the determined calibration scheme includes an interpolation scheme, a merging scheme, and an averaging scheme.

12. The method of claim 9 , wherein the mapping comprises mapping the segment vector to the battery state space based on a predetermined mapping parameter.

13. The method of claim 12 , wherein the predetermined mapping parameter comprises a predetermined reference matrix; and

the mapping comprises projecting the segment vector onto the predetermined reference matrix to extract the feature vector having a dimension corresponding to the battery state space.

14. The method of claim 13 , wherein the predetermined reference information comprises information on patterns for the battery state types in the battery state space; and

the estimating further comprises determining the state of the battery by comparing the feature vector to the information on the patterns.

15. The method of claim 14 , wherein the battery state types comprise a normal state type, an abnormal state type, and a fault state type; and

each of the abnormal state type and the fault state type comprises at least one subtype.

16. The method of claim 14 , wherein the determining comprises:

calculating a similarity between the feature vector and each of the patterns in the battery state space; and

determining the state of the battery based on a comparison between the calculated similarities.

17. A non-transitory computer-readable storage medium storing instructions for causing computing hardware to perform the method of claim 9 .