Study On The Model For The Determination Of The Time To Avoid Collision Based On Fuzzy Pattern Recognition Theory

Among all the marine traffic accidents, shipping collision ranks the first. Therefore, to avoid shipping collision or to decrease shipping collision has become a pressing task for the shipping industry. Since the 1950s and 1960s, the scholars in navigation field have attached great importance to the collision avoidance and gained fruitful achievements. In the studies of Ship collision avoidance decision-making, how to determine the time of collision avoidance is a concerned issue. The research on the time of collision avoidance will for one part give power to the development of the collision system, and for another part will provide reference for the ship drivers. This paper will first introduce the research of shipping collision avoidance and the analysis of parameters affecting the timing of avoidance. Then a decision model will be built based on the theory of fuzzy pattern recognition.This thesis mainly deals with the following aspects:1. Research on the effect of course changing. When the steering angle remains the same, the direction of relative movement is the reflection of avoidance effect. Use Geometrical plotting to build a mathematical model; render a 3D image. When the steering angle remains unchanged, the impact of steering angle, relative bearing of the coming boats and ratio of speed on changes in the direction of relative movement will be analyzed to research the effect of course changing.2. Parameters affecting shipping avoidance. Based on the factors affecting the timing of course changing, this paper focuses on four factors including effect of course changing, variation of DCPA, and descriptions of ships and length of ships. These four factors will be quantified.3. Using fuzzy pattern recognition theory to build decision-making model of timing of collision avoidance. The thesis will divide the timing of collision avoidance into five standard models, including "most early", "earlier", "early", "just in time", "too late". Those selected influence factors are used as recognition index to conduct pattern recognition and the decision-making model will input distance and range of TCPA. When distance and TCPA reach a certain threshold value, vessels should take immediate action to avoid collision.With respect to the specific encountered situations, the decision-making model of timing of collision avoidance is used to make decisions under various encountered situations. First, these decisions will be analyzed separately. Then these decisions will be compared with those decisions taken by experts. The result suggests that the decision-making model is feasible in practice.