An Intelligent Reactive Controller for Ocean Science Autonomous Underwater Vehicles

[9211914]特纳海洋科学在理解涉及全球变化的生物地球化学过程中起着至关重要的作用，远程自主水下航行器（lrauv）是海洋科学家最有前途的工具之一。然而，智能控制的最新技术尚未在实际物理设备中得到应用；在高度受限的环境中使用。本课题旨在研究远程水下机器人的嵌入式、智能、无功控制器的设计与构建问题。该项目被称为ORCA，将用于控制海洋系统工程实验室（MSELs）目前正在开发的远程AUV。ORCA将建立在智能控制的现有工作基础上，主要是利用反应计划的人工智能子领域。研究者之前在基于图式的推理（一种反应性计划方法）方面的工作将被计划、不确定性推理和智能控制方面的其他工作所补充。ORAC将能够创建并遵循任务计划，以实现用户（科学家）的目标，根据需要修改其计划并采取行动，以应对意外事件以及不完整知识、不确定性和与不断变化的世界互动的其他影响。

9211914 Turner Ocean science has a critical role in understanding the biogeochemical processes involved in global change, and long- range autonomous underwater vehicles (LRAUVs) are one of ocean scientists' most promising tools. However, the state of the art in intelligent control has not seen application in real physical devices; that which has, been used in highly-constrained environments. This research is aimed at studying the problem of designing and building an embedded, intelligent, reactive controller for long- range AUVs. This program, called ORCA, will be developed to control the Marine Systems Engineering Laboratory's (MSELs) long- range AUV, currently under development. ORCA will build on existing work in intelligent control, primarily drawing on the artificial intelligence subfield of reactive planning. The investigator's previous work in schema-based reasoning, a reactive planning method, will be augmented by other work in planning, reasoning under uncertainty, and intelligent control. ORAC will be able to create and follow mission plans to achieve users' (scientist') goals, revising its plans and taking actions as needed to cope with unanticipated events and the other effects of incomplete knowledge, uncertainty, and interacting with a changing world.