Autonomous mission management for an unmanned underwater vehicle

摘要

With the development of autonomous underwater vehicle technology, there is an increasing demand of its autonomy. The paper presents a distributed hierarchical Petri net model of mission control procedure to describe logical relations among mission level, task level and behavior level. The mission level takes on the genetic algorithm to plan, replanning the global path according to speed, energy and mission constraints in unstructured, partially unknown or hostile environments. The task coordination Petri net implements task schedules and discrete events handling strategies. The task coordination algorithm based on the RW discrete event supervisory system theory has been proposed to coordinate the pre-programmed tasks and the newly fired tasks according to internal and external events. The task coordination level may trigger mission replanning events to the mission level when it fails to coordinate tasks. Details are given of semi-physical virtual system applying to validate the feasibility and effectiveness of the architecture design. The simulation results show that the architecture makes it possible to make on-line mission replanning according to the new mission and task coordination in the terrain scanning mission by AUV.