Autonomy Infused Teleoperation with Application to BCI Manipulation

Abstract: Robot teleoperation systems face a common set of challenges including latency, low-dimensional user commands, and asymmetric control inputs. User control with Brain- Computer Interfaces (BCIs) exacerbates these problems through especially noisy and erratic low-dimensional motion commands due to the difficulty in decoding neural activity. We introduce a general framework to address these challenges through a combination of computer vision, user intent inference, and arbitration between the human input and autonomous control schemes. Adjustable levels of assistance allow the system to balance the operator’s capabilities and feelings of comfort and control while compensating for a task’s difficulty. We present experimental results demonstrating significant performance improvement using the shared-control assistance framework on adapted rehabilitation benchmarks with two subjects implanted with intracortical brain-computer interfaces controlling a seven degree-of-freedom robotic manipulator as a prosthetic. Our results further indicate that shared assistance mitigates perceived user difficulty and even enables successful performance on previously infeasible tasks. We showcase the extensibility of our architecture with applications to quality-of-life tasks such as opening a door, pouring liquids from containers, and manipulation with novel objects in densely cluttered environments.

Authors: Katharina Muelling, Arun Venkatraman, Jean-Sebastien Valois, John Downey, Jeffrey Weiss, Shervin Javdani, Martial Hebert, Andrew B. Schwartz, Jennifer L. Collinger, and J. Andrew (Drew) Bagnell

Online Bellman Residual Algorithms with Predictive Error Guarantees 

Abstract: We establish a connection between optimizing the Bellman Residual and worst case long-term predictive error. In the online learning framework, learning takes place over a sequence of trials with the goal of predicting a future discounted sum of rewards. Our analysis shows that, together with a stability assumption, any no-regret online learning algorithm that minimizes Bellman error ensures small prediction error. No statistical assumptions are made on the sequence of observations, which could be non-Markovian or even adversarial. Moreover, the analysis is independent of the particular form of function approximation and the particular (stable) no-regret approach taken. Our approach thus establishes a broad new family of provably sound algorithms for Bellman Residual-based learning and provides a generalization of previous worst-case result for minimizing predictive error. We investigate the potential advantages of some of this family both theoretically and empirically on benchmark problems.

Authors: Wen Sun and J. Andrew (Drew) Bagnell

Watch as our ARM-S system autonomously stacks blocks as demonstrated on a touch based GUI!

Congratulations to Stephane for successfully defending his thesis on “Interactive Learning for Sequential Decisions and Predictions”

A copy of Stephane’s thesis is available here

Congratulations to Dan for successfully defending his thesis on “Inference Machines: Parsing Scenes via Iterated Predictions”.

His thesis is available at http://www.cs.cmu.edu/~dmunoz/munoz_thesis_13.pdf.

The New York Times has covered an article on the ARM-S project where we perform the task of changing a tire autonomously:

http://www.nytimes.com/2013/03/30/science/making-robots-mimic-the-human-hand.html?_r=0

Title:
Activity Forecasting

Authors:
Kris Kitani, Brian Ziebart, Drew Bagnell and Martial Hebert

Abstract:
We address the task of inferring the future actions of people from noisy visual input. We denote this task activity forecasting. To achieve accurate activity forecasting, our approach models the effect of the physical environment on the choice of human actions. This is accomplished by the use of state-of-the-art semantic scene understanding combined with ideas from optimal control theory. Our unified model also integrates several other key elements of activity analysis, namely, destination forecasting, sequence smoothing and transfer learning. As proof-of-concept, we focus on the domain of trajectory-based activity analysis from visual input. Experimental results demonstrate that our model accurately predicts distributions over future actions of individuals. We show how the same techniques can improve the results of tracking algorithms by leveraging
information about likely goals and trajectories.

For more details, head to the Project Page

The BIRD MURI team has recently published this video to showcase it’s research on autonomous avoidance of trees by micro UAVs in forests.

For more information, head to the BIRD MURI page!

This year we welcome Dov Katz, working with Drew Bagnell and Tony Stentz on interactive perception for manipulation in unstructured environments, Nick Rhinehart as a research programmer working on the VMR and finder projects, Andreas Wendel, a visiting PhD student from the Graz University of Technology as a visiting scholar working on the BURD MURI project and Kumar Shaurya Shankar as short term research staff working on the BIRD MURI project.

Computational Rationalization: The Inverse Equilibrium Problem

Abstract:Modeling the purposeful behavior of imperfect agents from a small number of observations is a challenging task. When restricted to the single-agent decision-theoretic setting, inverse optimal control techniques assume that observed behavior is an approximately optimal solution to an unknown decision problem. These techniques learn a utility function that explains the example behavior and can then be used to accurately predict or imitate future behavior in similar observed or unobserved situations. In this work, we consider similar tasks in competitive and cooperative multi-agent domains. Here, unlike single-agent settings, a player cannot myopically maximize its reward — it must speculate on how the other agents may act to influence the game’s outcome. Employing the game-theoretic notion of regret and the principle of maximum entropy, we introduce a technique for predicting and generalizing behavior, as well as recovering a reward function in these domains.