Dr. Marjorie Skubic
University of Missouri at Columbia

Abstract

There is much evidence for the existence of an adaptive working memory system involving prefrontal areas in the primate brain. Such memory is closely tied to the learning and execution of tasks, as it contributes to decision-making capabilities by focusing on essential task information and discarding distractions. Our goal is to develop such a memory for robots to support the following:

· Focus attention on the most relevant features of the current task.

· Support generalization of tasks without explicitly programming the robot.

· Guide perceptual processes by limiting the perceptual search space.

· Provide a focused short-term memory to prevent the robot from being confused by occlusions

· Provide robust operation in the presence of distracting irrelevant events.

In this talk, we will discuss the motivation, approach, and status of a new NSF ITR project in which an adaptive working memory is investigated for robot control and learning. In the project, we will integrate an adaptive working memory structure into a robot in order to provide the embodiment necessary for exploring the issues of task learning. In turn, this leads to a complex but realistic system involving perceptual systems, actuators, reasoning, and short-term and long-term memory structures. In addition, we will discuss planned experiments intended to evaluate the utility of the adaptive working memory, through the use of vision-based object recognition and human-inspired spatial language.