Model Driven Autonomous System Demonstration and Experimentation Workbench

To achieve high-level strategic goals, DoD elements are investigating the potential of adding autonomy to unmanned assets. An autonomous system must be able to amend its"pre-loaded"plan by rapidly assessing the current situation, examining a number of possible outcomes, initiating a course of action and continuing to reassess its decisions and plan. Currently, data needed for training autonomous models, as well as response algorithms and heuristics, are limited for the system architect designing an autonomous platform. In areas where such data is available, such as observations of the recent DARPA Grand Challenge for autonomous automobiles, no well-developed system for identifying, archiving and maintaining useful information related to autonomous platform design exists. A well-defined framework for Autonomous Demonstration and Experimentation could collect real-world data, design decisions, requirements and experiment results into a scalable and intuitive workbench. This workbench will greatly aid the design of autonomous systems and identify gaps as well as provide management with actionable data.