Research On Plan Repair Method Of AUV Based On Behavior

Autonomous Underwater Vehicles(AUV)as an important tool for the development and utilization of marine resources,have the ability to autonomously perform a variety of tasks in dynamic marine environment.There are many kinds of uncertain events likely to encounter in the process of mission,Because of the uncertainty in marine environment.The uncertain events will not only affect the executed of tasks,but also affect the safety of AUV.When AUV in the face of uncertainty events,in order to successful completion of the given tasks,AUV must be replanning to make itself can adapt to the changed environment.This paper use a planning repair method to rapid repair the original plan to enhance the AUV intelligence level and adaptive level,when uncertain events occur in AUV behavior layer.Firstly,this paper introduces the function of each layer of the structure of AUV hierarchical planning system,every layer is the independent but contact closely.Through the study the characteristics of the uncertain events and the repair ability of the behavior layer,the GPS correction,navigation deviation,emergent obstacles and target losing four kinds of uncertain events were repaired in the behavior layer.Make the practical problem that need to be solved in the behavior layer planning repair clearly.Secondly,according to the research on planning repair algorithm,combined with the characteristic of behavior planning,real-time,small scale and high priority,the planning repair method based on insert and delete is determined,and improve insert part and delete part for application in behavior layer.Removal tree is used to determine the actions that must be deleted in delete part,and heuristic method is used to calculate the cost of remove tree,the tree that has smallest cost is removed priority.Planning method is used to find a new planning in the feasible solution space that produced by delete the remove tree in the insert part,achieved with the purpose of using effective part to replace failure parts.Finally,combined with the AUV actual demand and uncertain events,the planning repair method based on insert and delete apply to behavior layer to repair the navigation task and regional search task.Emergent obstacle event is processed in the navigation task mainly,the A* algorithm is improved by using the method of insert and delete,the improved A* algorithm can repair of emergent obstacles.The target lost event is processed by insert and delete task status in the regional search task.Summarize the general methods to make other tasks with fast repair capacity,based on the transformation of two tasks.The simulation experiment of navigation task,regional search task and pipeline tracing task prove the effectiveness of the method based on insert and delete in the dynamic marine environment for the uncertain events fast repair,prove the same uncertain event repair methods in different task events in general at the same time,easy to expand to other tasks.The planning repair method based on insert and delete make AUV behavior layer have fast repair ability,when the uncertain event occurs,improve AUV adaptive level in dynamic ocean environment.