Research On UUV Fault-tolerant Adaptive Task Planning And Decision-making Based On Unexpected Events

In recent years, as human continue to explore the marine environment, UUV (Unmanned Underwater Vehicle) technology has been rapid development. Because of the marine environment's complexity and unpredictability as well as real-time constraints of the underwater communication, higher requirements on the autonomy and robustness of UUV have been put forward. This paper basing on UUV as the research background, focuses on all kinds of unexpected situation existing in the marine environment to build research on UUV system structure and function module, the UUV semantic knowledge network establishing,unexpected event detection and planning, and other aspects. At last, we used simulation platform to verify the effectiveness of the proposed method.Firstly, through the analysis and comparison of UUV classical architecture, typical deliberative/reactive hybrid architecture has been chosen, and detailed design and division of functional modules have been made for the architecture, meanwhile the functional modules have been analyzed in detail.Secondly, relative knowledge about ontology theory and semantic knowledge network have been introduced. According to the establishment process of semantic knowledge network,UUV semantic knowledge network has been build based on ontology theory, making UUV builds logical relationship between each system to realize the knowledge sharing and reuse.Thereby UUV could make effective decisions through the reasoning ability of semantic knowledge network when it faces unexpected events.Thirdly, the paper focusing on the unexpected events that UUV may encounter during the execution of tasks makes detailed analysis and classification, and takes depth research of the unexpected events detection method in the ontology model of the basic of semantic knowledge network, and designs the unexpected events detection SWRL rules and carry on the explanation of the rules, as well as makes detail instruction aiming at the planning and decision-making process of unexpected events.Finally, the simulation experiment platform has been build and UUV task planning and decision-making simulation experiments under the condition of non event trigger and unexpected events have been designed. Through the analysis of the results of simulation experiment, the UUV unexpected events planning and decision-making capability based on semantic knowledge network has been verified.Through the establishment of UUV semantic knowledge network, achieve the interoperability of relative field information, and improve the cooperation capability,independent operating capability and situation awareness capability of each system within the platform, making UUV has higher flexibility,robustness and autonomy during the task execution process.