Research On Key Technologies Of Autonomous Recovery And Docking Of AUV Based On Deep-sea Operation Platform

In the context of the demand for large-scale exploration of marine resources in the future,the deep-sea operation platform,as the underwater docking base station of autonomous underwater vehicle(AUV),can provide energy and information exchange for AUV for a long time,greatly improve the use efficiency of AUV and reduce the cost of large-scale exploration and development.In order to ensure that AUV can return to the deep-sea operation platform independently and accurately,the path planning and precise guidance docking of AUV in the deep-sea environment become key technologies.Therefore,based on the deep-sea operation platform,this thesis deeply studies the path planning in the process of autonomous recovery of AUV in the remote stage and the relevant technology of accurate guidance to the docking station in the short-range stage,so as to realize the autonomous recovery and docking of AUV into the deep-sea operation platform,which is of great significance for the human to accelerate the understanding,development and protection of the ocean.This article specifically conducts research on the following parts:For the research of global path planning,this thesis proposes an AUV global path planning method based on the improved BAS algorithm.First,by analyzing the performance of various path planning algorithms,an improved optimization scheme is proposed for the limitation of BAS algorithm.Then,the planned global path is optimized through Bézier curve to make the optimized global path more consistent with the kinematics characteristics of AUV.Finally,establish a three-dimensional map model of the seabed using MATLAB simulation software and conduct simulation validation experiments.For the research of local path planning,this thesis analyzes the deep-sea obstacle avoidance environment of AUV,and proposes a dynamic obstacle avoidance method of AUV based on improved VO algorithm.Firstly,by further analyzing the obstacles,the speed obstacle domain in the algorithm is improved.Secondly,the real-time position information of obstacles collected is used to fit and predict the collision time,which is combined with the collision risk to determine the timing of starting and ending obstacle avoidance.Then,the collision avoidance rules are introduced into the underwater obstacle avoidance of AUVs to ensure their correctness.Finally,simulation validation experiments were conducted using MATLAB simulation software.For the accurate docking of AUV in the short range stage,this thesis proposes to conduct accurate docking through single light source vision guidance by analyzing the deep-sea recovery docking environment.Firstly,the method of short-range light source guidance is designed.Secondly,the optical characteristics of seawater are analyzed,and the LED light source color used to guide the light source in the seabed environment is determined.Then analyze the visual image processing technology,and process the image acquired by AUV during the guidance process.Finally,the relative position of the AUV docking process is estimated by the extracted image information,and the relative position data is fed back to the AUV control system to realize the accurate docking of the AUV.At the end of this thesis,the AUV hardware scheme and software scheme are designed according to the proposed AUV short-range light source guiding docking strategy,and the small AUV experimental prototype for experimental verification and the simulation dock cabin of the deep-sea operation platform are built.Then the feasibility of the single-light vision-guided docking scheme is verified by the pool experiment simulating the seabed environment.