Research On UUV Path Planning And Avoidance Method In Dynamic Obstacle Environments

Unmanned Underwater Vehicles(UUV)are affected by various obstacles,such as submerged reefs,islands,or passing ships,during the navigation process.The UUV’s autonomous collision avoidance and path planning capabilities are an important prerequisite to ensure that it can safely avoid various obstacles and navigate to the specified sea area to perform tasks.However,the uncertainty of forward-looking sonar’s detection of dynamic obstacle information is one of the important reasons for the insufficient collision avoidance ability of UUV.Therefore,this article mainly focuses on UUV path planning and collision avoidance methods in dynamic obstacle environments.The research content is as follows:Firstly,carry out the contour construction of static obstacles.The polar coordinate system is established according to the distance and azimuth information of the forward looking sonar,and the conversion between the polar coordinate system and the world coordinate system is completed.Using the support vector clustering method,the obstacle data collected by the forward-looking sonar is clustered to realize the construction of the obstacle outline and eliminate abnormal data points to ensure the accuracy and reliability of the data.Simulation results verify the effectiveness and robustness of the proposed support vector clustering method.Secondly,a dynamic obstacle movement model and a UUV three-degree-of-freedom kinematics model are established to estimate the movement state of the dynamic obstacle.According to the characteristics of UUV,combined with the knowledge of ship motion modeling,UUV kinematics model is established,and the motion model of different dynamic obstacles is analyzed.The variable-dimension Kalman filter algorithm is used to complete the position estimation and speed estimation of dynamic obstacles.Simulation results verify the effectiveness of the designed variable-dimensional Kalman filter algorithm.Thirdly,the UUV collision avoidance danger degree and fuzzy control algorithm are combined to avoid dynamic obstacles.The collision avoidance process and encounter situation between UUV and dynamic obstacles are analyzed,and corresponding methods for avoiding dynamic obstacles are proposed according to different collision avoidance scenarios.According to the motion state of different dynamic obstacles,the obstacle avoidance model is established and the UUV collision avoidance risk is studied.Through simulation experiments of UUV path planning and collision avoidance under dynamic obstacle environment,the feasibility and effectiveness of the proposed fuzzy control algorithm are verified.Finally,the improved artificial potential field method is used to avoid moving and static obstacles.Aiming at the target unreachable and local minimum problems in the traditional artificial potential field,the distance between the UUV and the desired target and the estimated collision time between the UUV and the obstacle are added to the repulsive potential field function.Using the improved artificial potential field method,the research on collision avoidance and path planning under the environment of dynamic obstacles and static obstacles coexisting is completed.Simulation results verify the feasibility and effectiveness of the improved artificial potential field method.