The Research Of Path Planning Based On ELM For AUV

Autonomous Underwater Vehicle, named AUV for short, is a kind of platformswith autonomous navigation and planning capabilities for detecting underwaterenvironment and developing ocean. It could take the place of human to work in thedangerous or inaccessible areas, which expands the underwater detection range andenhances underwater operation ability for human. Path planning has a vital role ofmovement to intelligent underwater robot, so it is a hot topic of research for a longtime. The AUV equipped with various kinds of sensors starts from the current locationand collects surrounding environment data to avoid obstacles safely, at last, reachesthe target by the optimal path, which is directly related to the safety of AUV andfinally related to the success of autonomous navigation.First, the Paper introduces the domestic and foreign current research situation ofAUV and analyzes the principle and the commonly used method of the path planningof AUV, what’s more, compares with the advantages and disadvantages of variousmethods and analyzes shortage of the methods to path planning. The path planningmust solves two problems: the environment modeling and path searching.In the global path planning, the obstacles known in environment map areexpanded and pretreated, then, made quadtree environment model, which saved thepreserved space of environment information and more quickly searched the path. TheAstar algorithm searched the path to make sure the way is the shortest one.Extreme Learning Machine, named ELM for short, is applied to the local pathplanning of AUV in the paper and it could make up the shortage of badgeneralization、slowly processing speed、falling into local extremum easily forother algorithms. The dynamic grid method, integrated the advantages of ELM, was to plan path for avoiding obstacles in the CV value was greater than1for local pathplanning. Finally, the process was simulated by using the MATLAB platform and thecharacteristics of the ELM was researched in path planning, what’s more, some resultshad been achieved through practice.