"This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5)."
In this project the PIs will investigate human-robot interaction in situations where a human is highly dependent upon a robot that serves as the medium for the outside world, such as emergency response, hostage negotiation, and healthcare. In these domains, the human "dependent" is typically connected to multiple other human "controllers" (medical specialists, structural engineers, rescue operations officials, etc.) via the robot proxy for long periods of time. The literature suggests that under these circumstances the dependent will respond to the robot socially, and will become distrustful as well as cognitively confused by a robot that presents a different affect for different controllers rather than a consistent communication strategy. The PIs believe that such a robot would occupy a novel "social medium" position within the Computers as Social Actors (CASA) model, and would be perceived as a loyal and helpful go-between who is an advocate for the dependent rather than as a device for accomplishing the goals of the controllers. To explore this hypothesis, the PIs will make use of the Survivor Buddy, a multimedia attachment for a robot which allows trapped victims to engage in two-way video conferencing, watch news, listen to music, etc. Formative experiments will be conducted at Stanford's CHIMe lab, followed by comprehensive, high fidelity experiments at Texas A&M's Disaster City using point-of-injury care scripts developed under prior work with medical doctors and rescue professionals. The paralinguistic aspects of the associated communication strategy will couple the ongoing work by Nass in voice characteristics and mannerisms with research in affective physical mannerisms in non-anthropomorphic robots under development by Murphy; the intellectual merit of the project thus stems from its multidisciplinary merging of communications and computer science. The research will introduce the CASA spectrum of relationships as a complement to robot-centric taxonomies, and will define a new relationship where a human is highly dependent upon a medium for long durations along with a new identity of social medium, which the PIs expect will have a greater impact on integrating robots into society than autonomous social actors and tele-operation. Project outcomes will include creation of a formal and comprehensive communication strategy for HRI, which combines verbal and nonverbal affect; this will unify the theory and practice of social robots, thereby breaking the pattern of ad hoc application of affect currently seen in the robotics literature and establishing the fundamental models and paradigms for continuing basic research in HRI.
Broader Impacts: This project will ultimately help save the lives of victims of accidents, disasters, and terrorism, and will also generally improve the quality of life for "shut-ins." The concept of social medium is well-matched to the current capabilities of tele-operation and semi-autonomy in civilian and military robotics; thus, the communication strategies developed will be immediately applicable to domains such as law enforcement and emergency response, which currently use robots, as well as to healthcare, where robots have not yet found a strong economic niche but have huge economic potential. The educational outreach plan includes multi-disciplinary curriculum development, as well as outreach to K-12 teachers and museums.
