Combining epistemic planning and belief-based programming for unified task and motion planning
This project will combine advances from epistemic planning (embodied in planners such as PKS) with high-level programming constructs enabling the representation of beliefs and stochastic information (such as the ALLEGRO language) for the purpose of unified task and motion planning. A new planning framework will be built, supporting the generation of robot plans that involve bridging the gap between reasoning about high-level task actions and low-level robot motion control. Plans will be built for dynamic real-world environments, such as those involving collaborative human-robot interaction or complex object interactions. Appropriate interfaces to these tools will be developed in ROS to integrate them with existing robot simulators and ecosystems.
A. Gaschler, R. Petrick. O. Khattib, and A. Knoll (2018). KABouM: Knowledge-Level Action and Boundary Geometry Motion Planner, JAIR 61:323-362.
V. Belle and H.J. Levesque (2015). ALLEGRO: Belief-Based Programming in Stochastic Dynamical Domains, IJCAI.