Temporal-Hierarchical Emergency-Degree Inference System for Running Vehicles Using Image and Navigation Data
Satoshi Koizumi*, Masayuki Matsushita*, Yasufumi Takama*, Hiroshi Takahashi** and Kaoru Hirota*
*Department of Computational Intelligence and Systems Science
Interdisciplinary Graduate School of Science and Engineering Tokyo Institute of Technology 4259 Nagatsuta, Midori-ku, Yokohama 226-8502, Japan
**Nissan Research Center, Nissan Motor Co., Ltd. 1 Natsusima, Yokosuka 237-8523, Japan
A fuzzy inference system of emergency-degrees for operating vehicles is proposed. Three situations of joining, cutting-in, and approaching the vehicle ahead are selected as target situations. The system has a hierarchical structure with temporal effects, using vagueness concept and a sigmoid function, to achieve the continuity of inference values and to suppress the influence of noise. The proposed system is tested using video data, taken by CCD stereo cameras mounted on a car and navigation data. The difference between the emergency-degree and driver’s perception of danger is evaluated by an expert for safety in an automobile company. Finally the validity of the fuzzy inference with temporal effects for the proposed system is shown through comparison with four other inference methods.