【作者】 韩丽洁；

【导师】 张汝波；

【作者基本信息】 哈尔滨工程大学 ， 计算机应用技术， 2010， 硕士

【摘要】 自主型水下机器人(AUV)能在非结构化的、不确定性的海洋环境下完成多项水下任务,任务规划起着关键的作用。任务规划不仅能根据下达的使命,规划出一个有效的任务序列,而且能对在水下任务执行过程中碰到的一定的不确定事件做出反应,实时的监督任务的执行。任务规划在一定程度上标志着AUV的智能水平,是进行决策和规划的关键组件。本文以AUV为研究背景,对大范围海洋环境下AUV的任务规划技术进行了系统的研究。首先明确了几项重要研究内容的提出,阐述了本文的研究目标,即保证水下机器人任务的顺利执行,以监督和管理的角度来对水下机器人进行任务规划。相对一般规划,本文的重点放在对整个任务的过程分析上,把任务规划作为一个下达任务,处理事件的监督与管理过程。考虑到AUV任务环境的特殊性,本文着重分析了AUV任务规划中可能会遇到的各种不确定性问题,并用智能规划的相关理论对这些不确定性事件进行了详细的研究,这些工作为以后的规划工作奠定了基础。AUV的良好体系结构是系统高效和合理工作的基础,任务规划在这个结构中的位置和作用是本文要考虑的首要问题。依据对任务规划的要求,本文给出了面向任务规划的体系结构的一个参考模型。当AUV的任务增加时,之前的单任务模型已经变得不是很有效。本文对任务组合的规划过程进行了建模研究,根据不确定性事件的发生情况,给出了任务规划过程模型并对其进行了仿真和状态分析,而且提出了重规划的方法,针对任务组合给出了一个简单规划软件,设定了任务的描述语言(格式),使得低一层结构能够解读其命令并指导规划的执行。对不确定性事件的分析和对任务规划过程的建模和仿真验证表明,本文提出的任务规划技术能够有效地处理任务组合中的各种意外事件,增强了任务规划过程中的可靠性和安全性,能指导AUV完成给定的多项水下任务。更多还原

【Abstract】 Autonomous Underwater Vehicle (AUV), which can complete various underwater tasks in the non-structured and uncertain ocean environment, benefits from the mission planning. Mission planning can give not only the effective order according to the mission assigned, but also react to the uncertain occurred in the mission process, monitor the mission implementation in real-time. Mission planning marks the level of vehicle to some extent and is the key assemble of the vehicle to planning and decision-making. This paper studies systematically the mission planning of AUV in a large sea environment.First of all, the paper clarifies several related research questions, and describes the goal of the research, which is to ensure the task successfully implemented and do the mission planning from the view of supervision and management. Different from general planning, this paper lays stress on the analysis of the whole process of the mission, treats mission planning as a supervision and control mechanism which can give the order and deal with uncertain events.The good architecture of AUV is a foundation to make sure the system efficiency and the work reasonable. The planning system’s position and function are the principal considerations. Therefore, according to the demands, this paper gives an architecture which can handle the mission planning.Considering the particularity of the task environment, this paper lays stress on the analysis of the uncertain questions in the process of mission planning. Using some related methods in AI planning, the paper describes the uncertain events in detail, and these lay the foundation of future work.The single task modeling loses its availability when the number of the task increases. The paper makes a research on the execution of the combination of main task, gives the modeling for the task to the performance of the uncertain events and the simulation and analysis. This paper provides some method of re-planning, give a simple planner to the mission combination and setting a kind of mission description language (format), to let the following executive model unscramble it and guiding the mission execution.The analysis of the uncertain events, the modeling and simulation of mission planning process show that, mission planning technology is able to deal with various unexpected events effectively in the mission combination, enhance the reliability and safety in the process of mission planning, and direct AUV to complete various given underwater missions.更多还原