Mission planning for autonomous marine vehicle operations normally considers journeys over long periods of time. Due to these long transit times, obstacles and hazards may be treated as independent. However, when docking, or operating in other similarly spatially constrained environments, this model may no longer be appropriate. In particular, the space between obstacles and the resultant constraints on vehicle motion may dominate the required planning model. This becomes important when considering the model of planning complex autonomous actions. Autonomous Surface vessels are increasing in capability to the point where they will be able to independently plan and execute complex tasks including many different spatially related subtasks. For an efficient selection of such tasks to be performed, the ability to explore and evaluate potential future states is required. To allow such exploration, the complex real-world environment must be reduced to a simplified model comprehensive enough to allow effective planning while still maintaining the simplicity required for effective search to be performed on future states. This paper discusses an alternate model for planning using a Delaunay triangulation for the creation of maritime planning domains in spatially constrained environments. Using data and tasks from the 2016 Maritime RobotX competition, this model is evaluated to assess the effectiveness and safety of the generated plans. The use of a Delaunay triangulation within this environment is shown to reliably generate mission plans that cover large and complex obstacle fields while selecting achievable paths and meeting task constraints.
- Intelligent robots, marine vehicles, path planning, unmanned autonomous vehicles