The study on intelligent vehicle collision-avoidance system with vision perception and fuzzy decision making

This paper proposes a combination scenario of vision perception and fuzzy decision making for developing an intelligent vehicle collision-avoidance system (IVCAS). In IVCAS, a CCD camera is installed on the following vehicle and used to capture the image of leading vehicles and road information. The features of the leading vehicles and lane boundary are recognized by vision perception method, which derived from our previous work on histogram-based color difference fuzzy c-means (HCDFCM). HCDFCM is a robust and fast algorithm for detecting object boundary. In this paper, we adopted the coordinate mapping relationship (CMR) with HCDFCM to provide a robust vision perception for the necessary information such as relative velocity, relative distance between leading and following vehicle and absolute velocity of following vehicle, etc. The collision-avoidance strategy is based on the vision perception and implemented by a fuzzy decision making mechanism. In this paper, the necessary information is integrated as a degree of exceeding safe-distance (DESD) to estimate the possibility of collision. A safety coefficient (SC) is defined to indicate the degree of safety. Therefore, the number of fuzzy rules that based on DESD and SC could be reduced to improve the efficiency of decision making. In addition to robust image processing, abundant information are derived from recognizing image feature using the proposed algorithm in this paper. The fuzzy decision making mechanism abstract useful compact data extracted from these abundant information. Therefore, the main advantage of IVCAS is using less number of fuzzy rules than other systems, and gets more effectiveness in vehicle collision-avoidance.