Learning over Subgoals for Efficient Navigation of Structured, Unknown Environments
[edit]
Proceedings of The 2nd Conference on Robot Learning, PMLR 87:213222, 2018.
Abstract
We propose a novel technique for efficiently navigating unknown environments over long horizons by learning to predict properties of unknown space. We generate a dynamic action set defined by the current map, factor the Bellman Equation in terms of these actions, and estimate terms, such as the probability that navigating beyond a particular subgoal will lead to a deadend, that are otherwise difficult to compute. Simulated agents navigating with our Learned Subgoal Planner in realworld floor plans demonstrate a 21% expected decrease in costtogo compared to standard optimistic planning techniques that rely on Dijkstra’s algorithm, and realworld agents show promising navigation performance as well.
Related Material


