| 摘要 |
The present invention provides a battery sensor for a vehicle. The battery sensor comprises: a precedent learning unit, a temperature sensing unit, and a season classification unit. The precedent learning part sets a plurality of pattern groups indicating a season; classifies a preset daily temperature data into each pattern group to set a self-organizing map used for learning; and generates a center coordination value of each pattern group on the self-organizing map as a previously learned season pattern data. The temperature sensing unit measures an external temperature data of the vehicle in real time. The season classifying unit receives an input of an outer temperature data and the previously learned pattern data; calculates a center coordinate value of the vehicle′s external temperature data using a k-means algorithm; indicates the calculated center coordinate value of the external temperature data on the self-organizing map to detect a pattern group with a center coordination value nearest to the center coordination value of the external temperature data; and classifies the season information indicated by the detected pattern group into a current season information. |
