University of Minnesota
University Relations
myU OneStop

Go to unit's home.

Home | DTI | 2004–05 funded proposals | William Schuler, Richard Voyles, Jeanette Gundel

Initiatives in Digital Technology: 2004–05 Funded Proposals

William Schuler, Richard Voyles, Jeanette Gundel

Multi-modal coaching of quasi-repetitive tasks

There are many domains in which it is desirable to allow casual users (non programming experts) to program robotic tasks. These include repetitive tasks in manufacturing, semi-repetitive tasks in loading/unloading and stacking in military, industrial, medical, and home applications, and non-repetitive tasks in home and healthcare applications. This document describes a proposal to build infrastructure for pursuing new funding in intuitive programming environments for robots, by combining existing complementary research strengths in human/machine communication. In particular, this proposed work will combine speech- and gesture-based modalities in a robot programming interface in order to allow users with limited formal training to quickly implement and refine unoptimized solutions for ‘quasi-repetitive' tasks like stacking or retrieving non-uniform objects. These kinds of tasks are broadly similar but not identical across iterations (as other manufacturing tasks are), and therefore require prohibitive programming costs to automate using conventional interfaces for dedicated programmers.