Nick Malone, Brandon Rohrer, Lydia Tapia, Ron Lumia, John Wood, "Implementation of an Embodied General Reinforcement Learner on a Serial Link Manipulator," IEEE International Conference on Robotics and Automation (ICRA), St. Paul, Minnesota, May 2012

Abstract— BECCA (a Brain-Emulating Cognition and Con-
trol Architecture software package) was developed in order
to perform general reinforcement learning, that is, to enable
unmodeled embodied systems operating in unstructured en-
vironments to perform unfamiliar tasks. It accomplishes this
through paired feature creation and reinforcement learning
algorithms. This paper describes an implementation of BECCA
on a seven Degree of Freedom (DoF) Barrett Whole Arm Ma-
nipulator (WAM) undergoing a series of experiments designed
to test BECCAs ability to adapt to the WAM hardware. In the
experiments, BECCA demonstrates 1) learning to transition the
WAM between states, 2) learning to perform at near optimal
levels on one, two and three dimensional navigation tasks, 3)
applying learning in simulation to hardware performance, and
4) learning under inconsistent, human-generated reward.