Leveraging people's everyday skill sets for interaction with robots

点击翻译按钮获取中文摘要

As advancing robotic technology begins to enter the everyday world of the general public, for example as utility or health-care robots, it is important to consider how people and robots will communicate, and how people will issue commands or extract data. Unfortunately, interacting with robots can be very difficult and generally requires training and expertise. As many robot users may not be robotics experts, I believe that it will be important to make interacting with robots as easy as interacting with other people. In this work we move away from robot-centric command consoles, and built robots that enable people to accomplish complex robot interaction tasks using their existing everyday (non-robot oriented) abilities and work practices (e.g., ability to teach others and give directions, or skills working with animals). We further aim to develop a stronger understanding of which existing social skills and structures robots can leverage in their designs, how they impact environments they integrate into, and how impact can be influenced by design. As one component of this work we will study which everyday tasks robots could assist with, and which existing interaction structures a robot could integrate into. Second, we will develop new robotic implementations which serve as proofs-of-concepts for incorporating existing human interaction techniques into robotic behaviors and designs. Finally, we will place our new technologies into real-world environments and observe people working with them as a means to test our designs and to further inform future interfaces. The resulting new human-robot interfaces will be among the first practical robot implementations that are designed to integrate naturally into existing everyday scenarios and social structures, and the in-context evaluations will provide new insight into how people interact with robots in real-world settings.