Applied Research Of Dynamic Collision Avoidance Based On Genetic Arithmetic For AUV

The research content of this paper is planning technology of collision avoidance for Autonomous Underwater Vehicle (AUV) based on information of a forward looking sonar. The main technology of the research is collision avoidance to moving objects based on genetic arithmetic. The technology includes two sides, motion estimation of objects and algorithms of collision avoidance.The paper researches the characteristics of collision avoidance sonar, design the mathematical model of collision avoidance sonar,and deal with the information of the obstacles.The paper researches real-time planning technology of collision avoidance and its strategy. Real-time planner of collision avoidance is designed in whole based on complex control framework of real-time collision avoidance for AUV, and five real-time behaviors of collision avoidance for AUV are designed. In order to realize effective collision avoidance moving objects, the paper researches prediction of moving objects based on the theory of object estimation and object prediction. Representative motion modes are proposed in the dissertation, standard kalman filter is used to predict the motion of non-mobile objects, self-adaptive kalman filter is used to predict the motion of mobile objects.AUV local collision avoidance system is researched and designed. AUV local planning framework of collision avoidance is constructed, collision avoidance strategy is based on shifting AUV's heading. The paper designs AUV collision avoidance arithmetic based on GA and narrates the correlative details how GA is applied to AUV collision avoidance. In order to improve the safety, availability and celerity, several actual problems are discussed which may affect AUV collision avoidance. At last simulations are done to validate the collision avoidance effects, the result shows that the method is efficiently.