Proc. of IEEE International Conference on Robotics and Automation (ICRA), New Orleans, April 2004. Initial Studies in Human-Robot-Human Interaction: Fitts’ Law for Two People Kyle Reed, Michael Peshkin, J. Edward Colgate, James Patton Laboratory for Intelligent Mechanical Systems Northwestern University Evanston, Illinois 60208 {reedkb, peshkin, colgate, j-patton}@northwestern.edu Often two people must work together physically on a common task, such as lifting and positioning a long board, or, in our model experimental system, turning a two-handled crank. Such tasks involve communication between the people, mediated by the task kinematics and dynamics: each person feels forces and motions produced by the other and derives some meaning from them. Tasks may include a degree of competition: the two people may not have exactly the same goal in mind, and must negotiate a compromise. Understanding human-human communication is important in designing robots for interaction with humans, and for robots that provide powered assistance for human-human tasks (such as physical therapy). In this paper we describe early experiments in human-human physical interaction, with a 1 dof robot included in order to give experimental access to the exchange of forces and motions between the people. We report on Fitts’ law-like tasks, in which the two people cooperate to move a cursor to a common target, or to targets that do not completely overlap. Our results suggest that human- human physical communication may be a rich area of study. Human-robot-human, human-human, haptic interaction, Fitts’ law I. MOTIVATION In numerous contexts, two people share physical tasks, and communicate through their physical interaction. Imagine two people working together to move a mattress through a doorway and up a flight of stairs. Each person is interacting with the other person through the object, each with his or her own forces and thoughts about where they are going. Similarly, imagine people exchanging a glass of water without spilling a drop of it. Some airplanes and helicopters have a flight stick that is coupled between the pilot and co-pilot. When both the pilot and co-pilot are holding the handles, they are working together and communicating through these handles. In each of these examples, physical cues are passed from one person to the other. It stands to reason that homo sapiens, a social species, has had opportunity to develop sophisticated ways of sharing a physical task, involving communicating with each other physically, compromising with respect to task goals, and teaching and learning manual skills. We expect that a significant channel of dyad communication should be through forces and motions, applied either directly to one another’s limbs, or via a mutually grasped object. Our motivating area of interest, lower limb rehabilitation, also involves haptic interaction between two people. The hands-on interaction between physical therapist and patient involves the communication of muscle tone, force, and motion to the patient, and the selective delivery of force and motion by the therapist. These examples suggest basic scientific questions in an area which we denote “human-robot-human” (HRH) interaction. We are exploring how two people physically cooperate, compromise, and guide one another, through force and motion, and how machine-generated forces and motions can enter into the human-human physical conversation. II. RELATED WORK Lifting and moving a table cannot be done individually, so people work together to accomplish this task. Takubo et al. [1] demonstrate how a robot can assist a person moving a table. The human is dominant and leads while the robot emulates a virtual non-holonomic constraint to keep it passively following. It is not known how two humans perform this same task. Learning how two humans communicate haptically through a rigid link will help in creating more intuitive communication between a robot and a human. Yanco and Drury [2] discuss many different combinations of how humans and robots can work together. All of their classifications are based on one or many humans giving an order to one or many robots and the robots doing the task. They do not classify the situation in which there are two humans giving an order to a robot that is acting on those same humans. Klingsport et al. [3] discuss various methods for communication between humans and robots, including programming commands, listening to natural words, and demonstration. They aim to make it as easy and intuitive as possible for people to work with a robot. To make it easy, human-robot communication in a shared task should follow the implicit human-human communication standards, which are poorly understood. Sallnas and Zhai [4] performed a study with a haptic system simulating the handoff of an object. They measured the time it