Abstract
Planning long duration missions for unmanned aerial vehicles (UAVs) in dynamic environments has proven to be a very challenging problem. Tactical UAVs must be able to reason about how to best accomplish mission objectives in the face of evolving mission conditions. Examples of UAV missions consist of executing multiple tasks such as: locating, identifying, and prosecuting targets; avoiding dynamic (i.e. pop-up) threats; geometric path planning with kinematic and dynamic constraints; and/or acting as a communication relay. The resulting planning problem is then one over a large and stochastic state space due to the size of the mission environment and the number of objects within that environment. The world state is also only partially observable due to sensor noise, and requires us to plan in the belief space, which is a probability distribution over all possible states. Some a priori contextual knowledge, like target and threat locations, is available via satellite imagery based maps. However, it is possible this will be "old" data by execution time. This makes classic approaches to a priori task, or symbolic, planning a poor choice of tool. In addition, task planners traditionally do not have methods for handling geometric planning problems as they focus on high level tasks. However, modern belief space geometric planning tools become intractable for large state spaces, such as ours. Recent tools in the domain of robotic manipulation have approached this problem by combining symbolic and geometric planning paradigms. One in particular, Hierarchical Planning-in-the-Now in belief space (BHPN) is a hierarchical planning technique that tightly couples geometric motion planning in belief spaces with symbolic task planning, providing a method for turning large-scale intractable belief space problems into smaller tractable ones.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have