AUV实时任务协调方法研究

自主地形勘察是AUV的一项重要使命,其使命控制过程是使命规划模块依据使命文本进行全局路径和任务规划,得到执行使命的任务序列.任务协调模块在预规划任务序列基础上,监控系统状态并触发事件,任务调度器根据事件消息对任务进行实时调度,发送行为控制指令驱动AUV动作来完成使命.给出了AUV任务协调模块Petri网模型,用来描述任务层离散事件间逻辑关系.在此基础上提出了基于改进RW理论的AUV实时任务协调方法进行任务间协调,实现AUV任务的实时调度.最后在虚拟仿真平台上结合地形勘察案例验证了任务调度算法,结果证明AUV能够在无人干预下自主地完成整个使命.

Study of AUV's real-time task coordination method

Terrain scanning is an important mission of AUVs (autonomous underwater vehicles). An AUV makes global path planning and task planning based on the mission text and forms the task lists. A task coordination module coordinates the preprogrammed task lists and the newly generated task, which is fired by event messages supervised by the state supervision system. Then the task coordination module sends behavior commands to drive the vehicle to finish the mission. Here, a task coordination Petri net model is presented to describe the event_based interactions at task level. A real_time task coordination algorithm based on the RW discrete event system theory is proposed to implement the real_time task coordination. The task coordination algorithm was tested according to a terrain scanning scenario in a virtual system and demonstrated that an AUV can fulfill its mission autonomously.