Research and application on information abstract method of data fusion

The characteristic of information fusion was analyzed and it was put forward that the key point of information fusion is the characteristic information fusion. Principal component analysis and rough sets theory were adopted on the information abstract for the oil analysis data of the power-shift steering transmission's (PSST) road test, and the key contamination particle size was found. Correlation analysis was adopted to estimate the pertinence between the characteristic information. The test proved that these two methods can be adopted on the information abstract and can get the accordant results, while correlation analysis can fetch up the two methods' insufficiency. This research result was applied on the contamination control of the PSST and the fault caused by oil contamination was prevented.