Research On Simultaneous Localization And Mapping And Path Planning For Unmanned Underwater Vehicle

UUV(Unmanned Underwater Vehicle)is an important tool for exploring complex underwater environment,it plays an important role in seabed environment observation,underwater salvage,ocean resource exploration,etc and has vast development space.However,because of the constraints of existing airmanship,flexibility and autonomy of UUV are greatly reduced in complex underwater environnent.Then SLAM(Simultaneous Localization and Mapping)and RRT(Rapidly Random Tree)realize overall environment map building,location and path planning,they become focus of researches on UUV.(1)Among existing SLAM algorithms,Extended Kalman Filter SLAM algorithm has a better precision at map building.However,when the data used to build a map is bigger,it follows more algorithm workload,longer calculation time and poorer efficiency.To solve the huge calculation problem of EKF-SLAM,this paper improves existing EKF-SLAM algorithm.It firstly uses sparse environment descriptions to build SLAM feature map,then uses rasterization map to embody the map and lastly proves this method has a good efficiency and accuracy by carrying out experiments.(2)Among existing path planning algorithms,RRT(Rapidly-exploring Random Tree)algorithim is more suitable for nonholonomic motion planning and multi-degree robot motion planning.Because of the slow convergence rate and lack of stability of RRT algorithm,planning path may be incomplete and unreliable in the path between start point and end point under known circumstances.Aim at this problem,this paper quotes an improved RRT algorithm-RRT with Auxiliary Path(RRT-Path),and proves improved path planning algorithm having better feasibility and stability by simulation experiment on the two different RRT algorithms.