Fuzzy K-nearest Neighbor and its Application to Recognize of the Driving Environment

Recently, some applications of information technology (IT) are studied in the world. In the automobile industry, the intelligent car is developed as one of applications of IT. The development of the intelligent car is very important and the recognition of the driving environment is the basic technique in it. In this paper, we try to recognize the driving environment by two supervised classification techniques, if-nearest neighbor (KNN) and fuzzy if-nearest neighbor (FKNN). KNN is a basic technique of supervised classification. FKNN has been proposed by one of the authors and it is a extension of KNN to introduce the concept of fuzzy theory. To compare with KNN, FKNN has the following advantages. (1) The range of K of FKNN is wider than KNN. (2) In case to use similarity based on metric, the results of KNN don't depend on the similarity. On the other hand, the results of FKNN depend on it. In other words, it is easier to tune up FKNN than KNN. The usefulness of FKNN is verified through the application to the problem.