中文摘要：

水面无人艇（ USV）是一种重要的海洋自主机器人，在障碍物环境中自主航行问题是当前USV的重要研究内容。针对高速USV在动态环境中的危险规避问题，提出一种基于行为的动态危险规避算法。算法首先对USV的运动特性进行分析获得基本运动空间，采用碰撞锥理论对USV与障碍物之间的情况进行判定，将海事规则约束和碰撞约束转换为USV基于行为的约束，通过求解基于偏航角度和速度的优化问题获得USV最优规避行为。仿真实验结果证明所提出的算法能够有效引导USV在高速（30 kn）情况下对动态障碍物实现有效危险规避。

英文摘要：

An unmanned surface vehicle （ USV） is a kind of important marine robot , and the problem of the autono-mous navigation in the obstacle environments is the significant research issues of USV .In this paper , an algorithm of dynamic obstacle avoidance based on behavior is proposed for the navigation of high -speed USV in the dynamic obstacle environments .Firstly, the basic motion space is calculated by analyzing the motion characteristics of USV , and according to the theory of collision cone the collision states of USV can be determined .The constraints of Inter-national Regulations for Preventing Collisions at Sea （ COLREGS） and collision of USV are transferred into the be-havior based constraints , and then the optimal avoidance behavior can be obtained by solving the yaw-angle and ve-locity based constraints optimization problem .In the end, the validity of the algorithm is demonstrated by high-speed （30 kn） USV in the dynamic obstacles environments .