The field of robotics is rapidly approaching a paradigm shift in how human beings collaborate, program, monitor, correct and generally interact with robotic systems. Several factors are driving this paradigm shift. First, robotic systems are becoming more capable: The capability of a robot has expanded from limited movement and grasping to complex sensing using force and vision and advanced manipulation. Many new systems are also safe for physical interaction. Along with improvements in hardware have come drastic advances in algorithms for motion planning, object detection and other perception, manipulation and human tracking.
This increase in capability has allowed robots to be used for more and more complex tasks, and in a wider range of use domains. Initially almost all robots were used for large scale industrial manufacturing, where they predominantly saw use for structured tasks such as material handling, welding and painting. However as robotic systems become more capable and expand out of the structured, interaction-free world of large scale manufacturing, and into the unstructured human-centric environments of small scale manufacturing and in-home assistance, we must re-think how humans interact with these systems.
We present several methods for improving human instruction, collaboration and general interaction with robotic systems by following the model of human apprenticeship. We show that such interaction between novice human users and robots is possible by formal modeling of the capabilities of the robot, providing specific information to these capabilities to make them task relevant, and providing representation for building a task from these capabilities in a manner that enables the generalization and reuse of task information. We further use advances in Virtual Reality to provide a dynamic and intuitive robot interaction and instruction environment.
Speaker Biography
Kelleher Guerin received his Bachelors of Science in Optical Engineering from the University of Arizona, and a Masters of Science and Engineering in Electrical Engineering from Carnegie Mellon University. His work has included the design of optical systems for space applications, robot visions for lunar excursion, robotic mining applications, large-scale wall displays, and robotic systems for small manufacturers. He is focused primarily on problems in Human-Robot Interaction, Human-Machine Interaction and Robotics for Manufacturing.