Research On Collision Avoidance Technology Of Unmanned Surface Vehicle Based On Maritime Rules

Unmanned Surface Vehicle(USV) is an important water unmanned platforms. As an important future naval combat platforms under information-age conditions,USV has been paid much attention by the navy in different countries. Therefore the U.S Navy specially formulated "Navy unmanned surface boats plan" as a guide for unmanned surface boats developing. USV usually works in the complexity of environment and with high speed, which is used to perform tasks in dangerous waters and not suitable for the task performed by someone warship and vessels. Automatic control technology as the core part of USV, which is the key technology to realize USV independent movement and autonomous task execution.Because the USV monement environment is complex, dynamic and unknown, so how to realize the obstacle avoidance and security has become the focus of the study.USV usually needs navigation safety in a highly dynamic and unknown environment, so autonomy must achieve a high level. Traditional risk avoidance technology is mainly applied in the low speed robot, which has been unable to meet the motion characteristics of USV.Based on the rules of the international maritime(COLREGS), a method of collision avoidance is proposed, the subordinative function and weighted method are used to calculate collision risk index. Through the collision risk index distinguish the current environment and establish a rapid feedback mechanism, to control the USV and complete the obstacle avoidance.Research on the COLREGS as the foundation, through the analysis and extraction of relevant circumvention rules, to build a complete USV risk aversion of knowledge base and establish a reasonable inference process knowledge as the USV of the intelligent decision support system data model library. Initial construction of the knowledge base is short of knowledge and expert support, there are redundant and conflict issues in knowledge, which seriously affect the efficiency of intelligent decision-making system. In order to solve the problem, attribute reduction in rough set and the theory of multiple attribute decision making are used to complete the reduction of attributes of the knowledge base and optimization of rule. Finally, a method of collision avoidance based oncollision risk index and multi-objective risk of circumvention technology based on local optimization are described, which is used to solve single objective and multi-objective dangerous situation of avoidance.In the simulation experiment, research set up a simulation environment for USV, which achieve the operating environment and sensor data as well as the number of states for motion simulate. Through the matlab software realize the data analysis for USV circumvent process,and then verify that the USV dangerous circumvention of technological feasibility and effectiveness. Simulation results and data analysis show that the method not only make USV follow the COLREGS , but also make USV achieve collision avoidance for the dynamic and static obstacles.