摘要: Current autonomous vehicle interoperability is limited by vehicle- specific data formats and support systems. Until a standardized approach to autonomous vehicle command and control is adopted, true interoperability will remain el... 展开 Current autonomous vehicle interoperability is limited by vehicle- specific data formats and support systems. Until a standardized approach to autonomous vehicle command and control is adopted, true interoperability will remain elusive. This work explores the applicability of a data model supporting arbitrary vehicles using the Extensible Markup Language (XML). An exemplar, the Autonomous Vehicle Command Language (AVCL), encapsulates behavior-scripted mission definition, goalbased mission definition, inter-vehicle communication, and mission results. Broad applicability is obtained through the development of a behavior set capturing arbitrary vehicle activities, and automated conversion of AVCL to and from vehicle-specific formats. The former uses task-level behaviors suitable for mission scripting and goal decomposition. Translations use the Extensible Stylesheet Language for Transformation, XML data binding, context-free language parsing, and artificial intelligence machine learning and search techniques. Translation capability is demonstrated through mappings of AVCL to and from multiple vehicle-specific formats. A final demonstration of the power of a common autonomous vehicle data model is provided by the implementation of a hybrid control architecture. The model's vehicle- independence and the ability to generate vehicle-specific data are leveraged in the design of an architecture that provides increased autonomy by augmenting a vehicle s existing controller. The utility of this architecture is demonstrated through implementation on the Naval Postgraduate School s ARIES Unmanned Underwater Vehicle. 收起