Event Location: NSH 3305
Bio: Debadeepta Dey is a 3rd year Phd student in The Robotics Institute, Carnegie Mellon University advised by Prof J. Andrew Bagnell. From 2007-2010 he was research staff in Prof. Sanjiv Singh’s group at the Field Robotics Center. He has worked on vision-based sense-and-avoid for UAVs, automated drilling for mining, robotics in agriculture and vision-based localization for heterogeneous robot teams. His main interests include bridging the gap between control and perception for autonomous mobile robots by developing new machine learning tools. Currently he is keen to make small and medium UAVs fly fast through dense obstacles using only vision.

Abstract: A popular approach to high dimensional control problems in robotics uses a library of candidate “maneuvers” or “trajectories”. The library is either evaluated on a fixed number of candidate choices at runtime (e.g. path set selection for planning) or by iterating through a sequence of feasible choices until success is achieved (e.g. grasp selection). The performance of the library relies heavily on the content and order of the sequence of candidates. Previous work in sequence optimization produces a static ordering. We propose a provably efficient method to optimize such libraries leveraging recent advances in optimizing submodular functions of sequences. This approach is demonstrated on two important problems: mobile robot navigation and manipulator grasp set selection. In the first case, performance can be improved by choosing a subset of candidates which optimizes the metric under consideration (cost of traversal). In the second case, performance can be optimized by minimizing the depth the list is searched before a successful candidate is found. Our method can be used in both online and batch settings with provable performance guarantees, and can be run in an anytime manner to handle real-time constraints.

In the second part of the talk, I will show an extension that yields a general approach to order the items within the sequence based on the context (e.g., perceptual information, environment description, and goals). We take a simple, efficient, reduction-based approach where the choice and order of the items is established by repeatedly learning simple classifiers or regressors for each “slot” in the sequence. Finally we demonstrate the efficacy of the approach on local trajectory optimization techniques.