Dobson, Andrew. Compact representations for efficient robot motion planning with formal guarantees. Retrieved from https://doi.org/doi:10.7282/T35X2D2K
DescriptionThis work provides compact representations for single- and multi-robot motion planning in the context of prehensile robot manipulation. This work describes the asymptotic near-optimality and probabilistic near-optimality properties of these planners. Probabilistic near-optimality is leveraged to provide practical and grounded stopping criteria for these methods which probabilistically guarantee the methods return high-quality paths. It is also shown how these methods can be leveraged to produce a compact planning representation, which is a lightweight structure that is quick to query and easy to store. The work also outlines a compact representation for solving multi-arm manipulation tasks, and integrates a scalable, asymptotically optimal multi-robot motion planning method to provide scalable, globally asymptotically optimal task and motion planning in manipulation domains.